Working Profiles and Employment Regimes in Europe

Muffels, Ruud J. A.; Fouarge, Didier J. A. G. Article Working Profiles and Employment Regimes in Europe Schmollers Jahrbuch – Zeitschrift für Wirtschaftsund Sozialwissenschaften. Journal of Applied Social Science Studies Provided in Cooperation with: Duncker & Humblot, Berlin Suggested Citation: Muffels, Ruud J. A.; Fouarge, Didier J. A. G. (2002) : Working Profiles and Employment Regimes in Europe, Schmollers Jahrbuch – Zeitschrift für Wirtschaftsund Sozialwissenschaften. Journal of Applied Social Science Studies, ISSN 1865-5742, Duncker & Humblot, Berlin, Vol. 122, Iss. 1, pp. 85-109, https://doi.org/10.3790/schm.122.1.85 This Version is available at: https://hdl.handle.net/10419/292007 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Schmollers Jahrbuch 122 (2002), 85-110 Duncker & Humblot, Berlin Working Profiles and Employment Regimes in Europe By Ruud J. A. Muffels and Didier J. A. G. Fouarge Abstract For a larger portion of society, working life becomes a continuing sequence of short employment and unemployment spells. The aim of this empirical paper using three waves of the European Community Household Panel (1994-1996) is to acquire a deeper insight into the flows on the labour market and the factors that might be responsible for the great variations in these employment patterns between individuals and households both within and across countries. The focus will be especially on longitudinal patterns of employment according to the attachment of the person to the labour market in a given period of time. These longitudinal employment patterns are called working profiles and it is claimed that these profiles provide a better insight into modern labour markets because of their focus on the dynamics or changes of employment and unemployment status over time. We use an amended version of Esping-Andersen's welfare state typology to test whether longitudinal employment patterns differ across these regimes according to the typical institutional set-ups of their labour market and social security policies. We added a Southern regime type because we found that the classification including such a regime type performs better in explaining differences in longitudinal employment patterns across Europe than one without it. Within a job search theoretical framework we estimated some multinomial logit models to test whether indeed regime types matter in explaining transitions in longitudinal employment statuses over time. The results show that regimes matter and that the Southern regime performs markedly different compared to the other three regime types. JEL-Classification: J 64, C 23 1. Working Profiles and Changing Concepts of Work1 The concepts of work are changing. Life-time employment, meaning working 40 hours a week for 50 weeks a year over a 40-year period often 1 The analyses in this paper were carried out as part of a research project for the European Panel Analysis Group (EPAG; http://www.iser.essex.ac.uk/epag). Funding by the European Commission in the framework of the TSER program 'The Panel Project; European Panel Analysis' (SOE2.CT96.3023) is greatly acknowledged. We are grateful to Richard Berthoud and two anonymous referees for their comments. Schmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 86 Ruud J. A. Muffels and Didier J. A. G. Fouarge with the same or a limited number of employers, is losing ground in favour of more flexible and diverse patterns of life-time employment. European economic and monetary integration will give a further impetus to the upswing of the process of increasing mobility and turnover on the labour market. The rise of atypical work and non-standard contracts reflects the need of firms to adapt quickly to rapidly changing market conditions. These market changes are emerging particularly from developments in information and communication technologies. The notion of the 'flexible firm' implies that the adaptability of firms as far as personnel management is concerned, is partly also established by the creation of core and peripheral workers segments. The existence of these two segments does not necessarily imply that there is no or little mobility between them. On the contrary, for theoretical as well as empirical reasons it might be true that the mobility within and between the two segments is rather high and, therefore, overall labour turnover might even rise. Images of segmentation and dynamics are two sides of the same coin. Some individuals within the peripheral group of workers might move quickly into permanent jobs while others keep wandering around in the lower strata of the labour market. But neither are core workers deemed to stay in the same jobs; on the contrary 'job-hopping' might be a better strategy to raise one's long-term career prospects or permanent income than staying with the same employer. Quite a few individuals belonging to the better strata of the labour market might experience rapid moves from one job to another either within the firm or in the external labour market. Whether this is actually the case for all workers might be questionable, but the image of a world with large segmentation and hardly any mobility is far from reality. The concept of a 'transitional labour market' might gain importance in current labour markets. This implies that for a rather large portion of society, working life becomes a continuing sequence of short employment and unemployment spells. Lifetime employment with the same employer, although still the reality for many workers, will increasingly be unattainable for many new entrants to the labour market. This leaves aside the fact that modern workers are likely to be unwilling to stay with the same firm for their entire career due to changes in labour and leisure preferences (Muffels, 2000). The aim of this paper is to acquire a deeper insight into these flows on the labour market and the factors that might be responsible for the great variations in these employment patterns between individuals and households both within and across countries. The focus will be especially on patterns of partial employment according to the attachment of the person to the labour market in a given period of time. These employment patterns are called Schmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 Working Profiles and Employment Regimes in Europe 87 working profiles and it is claimed that these profiles provide a better insight into modern labour markets because of their focus on the dynamics or changes of employment and unemployment status over time. The attachment to the labour market over a certain time and particularly the length of the intermittent periods of withdrawal from the labour market determine the extent to which people are in full or partial employment and therewith, whether they are fully or partially included. Then, in a dynamic perspective, labour market exclusion refers to longitudinal employment patterns over time with intermittent periods of unemployment, disability or, even, early retirement. These intermittent periods of withdrawal constitute cases of exclusion only in so far as these are enforced due to collective layoffs, disability or retirement. If they are the result of free will, as might be true for temporary jobs or part-time employment, it is not justified to speak of exclusion because people prefer these jobs for their shorter working hours or for rendering more leisure time. With increasing age, in many cases a forced retreat from the labour market can be observed due to retirement, redundancy or collective layoffs. Therefore, labour market exclusion in terms of enforced idleness might apply to the long-term unemployed, those seeking work but not able to get the job they want, the underemployed who work less than they want and the disabled and senior workers who are not expected to work. However, there is little information in the data of the European Community Household Panel about whether the exclusion is due to enforced idleness or free will and preferences for leisure.2 Especially for female and older workers this neglect of the unforced or forced nature of exclusion might lead to an over-estimation of the extent of exclusion. There is not much we can do here except to be cautious about drawing far-reaching conclusions on the extent of labour market exclusion. Therefore, when we use the term 'exclusion' it is in an non-normative, positive sense, meaning that the excluded do not take part in the labour market, without suggesting that it is enforced idleness that makes them 'excluded'. Data As employment patterns are affected by the design of the institutional setting in each country and by the process of European integration, single country studies and 'snap-shot' comparisons will not suffice. Therefore, we decided to use the European Community Household Panel (ECHP) covering 2 There is some information on the number of hours people want to work for the reference week, the week just before the date of interview, but not on the number of annual hours people want to work. However, that is the kind of information needed if the idea is to use the monthly calendar information to determine longitudinal work profiles. Schmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 88 Ruud J. A. Muff els and Didier J. A. G. Fouarge all EU countries except Sweden. The ECHP is an attempt to collect socioeconomic information on income, labour market and demographic changes, applying common definitions, data collection methods and data editing and cleaning procedures. We used the first three waves from 1994 to 1996. Hence, the information about transitions on the labour market captures 36 months of observation. At the European level, three years seem sufficient to observe a substantial number of transitions from steady jobs into marginal employment and vice versa, and from one longitudinal employment status into the other. Currently, four waves of the ECHP have become available covering the years 1994 to 1997. The four-wave dataset contains information on all EU countries including some Swedish data. Definitions of Longitudinal (Unemployment Considering the relatively short observation period in the mid-nineties, it cannot be ruled out that the mobility patterns we observe are influenced by differences in the business cycle across countries. At the time of observation, however, all countries under scrutiny were in the same (upward) phase of the business cycle. The focus in this paper is on longitudinal employment and unemployment patterns. The ECHP provides information on the activity status of the respondent for each month in the calendar year prior to the interview date. The activity status variable is then transferred to a longitudinal activity status variable called 'main status', i.e. the most frequent status observed on a monthly basis over the 36 months of observation. The frequency of monthly status, then, determines whether people are employed, unemployed or inactive over the 36-month period (usual status). In addition to the 'usual status', everyone whose main status is 'not working' in a particular month has been treated as being unemployed. Therefore, in this approach people are considered inactive only when they were not working during all of the 36month period, i.e. not working all of the time. 2. Unit of Analysis The unit of analysis is the individual within the household. The research subject involves the unemployment profile of the person but it is assumed that this individual profile is affected by the social and economic situation of the household, which, in turn, means the labour market and financial situation of the individual household members. Implicitly, it is therewith assumed that although individualisation patterns are widespread, the houseSchmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 Working Profiles and Employment Regimes in Europe 89 hold is still the economic unit within which important decisions with respect to labour supply are jointly made. In our empirical models, information on the household is, therefore, included to take account of the role of the family with regard to employment decisions. Only the adults of working age are considered in the analyses. Work insecurity has been defined as a weak attachment to the labour market in a longitudinal sense (working less than 50% of the potential 36month working period). Because the observation period is rather short, only short work patterns can be observed. If the time horizon could be extended, working patterns might change. To say that if people are working less than 50% of the potential working time over a three-year period they are work-insecure does not imply that they consider themselves to be, or feel, work-insecure. The subjective worksecurity status also depends on their preferences for work; i.e. the number of months and hours they actually want to work. In the ECHP there is some information on whether people are looking for a job and for how many hours. It is not known how many months during a year they really want to work and for what reasons they want to work less than the full 12 months. In future research the issue of how these labour supply variables might affect the outcomes of our analyses will certainly be tackled. 3. Employment Regimes and Flexible Labour Comparative research into the labour market should take account of the evidence that stocks and flows on the labour market are affected not only by the demographic and economic situation at the country level but also by cross-national institutional differences reflected in labour market policies and social security designs. Looking at the national settings it appears that there is great variety in goals, objectives, tools, institutions and policies. Despite this variety, the idea that these systems cluster one way or another in a limited set of welfare and employment regimes is well-known in the literature. These regimes represent different 'worlds of welfare capitalism' (Esping-Andersen, 1990, 1996, 1999), each being internally tightly integrated, and each being sharply differentiated from one another (Goodin et al., 1999). Each welfare state is, of course, uniquely defined by its own logic in terms of institutional set-up, policy design, and functioning but as suggested by Esping-Andersen also clusters around some distinct 'ideal-typical' regime type. His typology was criticised by authors like Leibfried (1992), Ferrera (1996) and Bonoli (1997) for his neglect of what they called a Southern or 'Latin-Rim' model of the welfare state. They argued that the Southern, Mediterranean countries belong to a different welfare regime Schmollers Jahrbuch 122 (2002)1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 90 Ruud J. A. Muffels and Didier J. A. G. Fouarge type with its familial characteristics and its immature and selective social security system granting poor benefits and lacking a guaranteed minimum benefit system. Esping-Andersen admitted in his later work (1996) that the Southern countries share some Catholic and familial traditions but do not form a specific type or group of countries, but were merely underdeveloped forms of the traditional Corporatist type (Arts and Gelissen, 1999). This issue will not be elaborated further here, since the data limitations confine us within the level of detailed analysis. The ECHP data cover transitional data for a period of three years only and since a number of country samples are small in size like Luxembourg, the Netherlands, Belgium, Greece and Ireland, there is a lack of sufficient transitions to conduct detailed analyses. Since the ECHP data contain information on the Southern European countries, it is feasible to test whether the Southern labour markets show an essentially distinct pattern. One could opt for a regional classification into Northern, Central European and Southern European countries. Apart from the lack of theoretical underpinnings for such a distinction, its disadvantage is also that the UK, Ireland and the Nordic countries like Denmark and Finland would be classified under the same heading. According to Esping-Andersen's classification of liberal, conservative-corporatist and social-democratic countries (1990) the UK and Ireland, as liberal welfare states in an 'ideal-typical'sense, should be set apart from the Nordic (Denmark and Finland; the ECHP does not contain any information on Norway or Sweden yet) and continental social-democratic welfare states (Netherlands). The classification of Ireland under the liberal heading by EspingAndersen (1990, 1999) is, however, rather doubtful. Considering a range of labour market indicators, it only shares the liberal feature of a low level of employment protection regulation. Looking at the other labour market indicators it seems to share the corporatist feature of an active labour market policy and the corporatist 'breadwinner's state' characteristic of a low female employment rate. In terms of familial characteristics, it shares the typical features of a Southern welfare state. Ireland should, therefore, be considered as part of a hybrid type of welfare state that does not fit into any of the 'ideal-typical' welfare states. In order to avoid the inclusion of Ireland as the only example of a hybrid type it was decided to keep it under the same liberal heading as the UK and to test, using the three-wave European panel-data, whether that makes sense empirically. For that reason it was decided to use an amended version of Esping-Andersen's classification. The UK and Ireland remain under the liberal heading, notwithstanding our reservations for Ireland, but the Southern welfare states are set apart as a distinct regime (Arts and Gelissen, 1999; Goodin et al., 1999). Countries like Germany, Belgium, France, Austria and Luxembourg belong to a continental corporatist type of welfare state and the Netherlands and the Nordic Schmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 Working Profiles and Employment Regimes in Europe 91 welfare regimes are classified under the social-democratic regime heading. The Southern regime cluster includes Spain, Portugal, Greece and Italy.3 Since our analyses involve the labour market performance of these welfare regimes, we will use the term employment regime instead of welfare regime. 4. Temporary Employment and Flexible Labour Contracts There is some evidence (European Commission, 1999; OECD, 1999) that the Southern labour markets are indeed quite different from the labour markets in Northern and Central Europe in some respects (employment regulation, unemployment level, share of non-standard jobs or flexible labour, share of informal economy). This can be shown when the distribution of non-standard and temporary employment is looked at across a number of European countries.4 In Table 1 the figures on the distribution of temporary, casual and other types of non-standard employment are depicted for 1996, the latest year for which we have information. Looking at the figures for non-standard, flexible employment it is not surprising that the largest share of flexible labour is found in countries with the highest level of employment protection regulation i.e. the Southern countries like Spain, Portugal, Greece and, to a lesser extent, Italy. Remarkably, though, Ireland also has a high level of flexible employment despite a low level of employment regulation. Smaller but still sizeable figures are found in the Northern countries, such as Denmark and Finland. Much smaller numbers are found in Luxembourg.5 Countries like France, the Netherlands and the United Kingdom, have sizeable levels of flexible employment, which are in between the numbers for the Southern and the Northern European countries. Nevertheless, they have average or low levels of employment regulation. The Northern countries, in common with some Southern countries, share the fact that 'casual work and other arrangements' are quite large. In Italy and Portugal these two job types account for 40% of all flexible jobs and in Greece even for 70%. In the United Kingdom and Ire3 To test the reliability of this classification, the empirical model for partial exclusion in Section 6 (see Table 6) has been estimated using country dummies instead of 'employment regime' dummies. It emerges that the model including the employment regime dummies 'captures' 97 % of the variance of the model with country dummies. 4 The Netherlands is included in the analyses although the Dutch data do not contain calendar information on the employment statuses over time on a monthly basis. However, using the information on the number of months people are in particular statuses it appeared feasible from these 'count data' to construct the longitudinal employment status variable indicating the number of months people are in employment or unemployment over the year (by giving precedence to work over unemployment and inactivity). 5 Figures for Luxembourg should be taken with caution because the number of observations is extremely small. Schmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 92 Ruud J. A. Muffels and Didier J. A. G. Fouarge land its share is about 50% and in the Netherlands 70% (particularly temporary agency and on-call contracts). It emerges that, except for Spain, countries with large numbers of atypical jobs have large numbers of these very 'typical' jobs like temporary agency work, on-call contracts, zero-hours contracts, labour pool jobs, freelance jobs, housework and the like. The welfare state classification does not provide much added value to the figures for the different countries. It appears that the liberal, social-democratic and corporatist regime types are very similar in terms of the relative size of the flexible workforce, whereas we would expect the highest level in corporatist and social-democratic countries. The highest levels are, indeed, found in the Southern regimes in which one in five employees has a flexible job. In Table 1 the figures from the ECHP 1996 are also compared with the Labour Force Survey (LFS) 1996 figures.6 The picture that comes out of the ECHP is not very dissimilar to the one that comes out of the LFS. A rather broad definition of flexible labour is also used in the LFS, including temporary jobs, temporary agency work, casual work7 and the like. The LFS figures provide identical rankings except for a few countries where remarkable and unexpected differences occur, such as for Ireland and Greece. The countries showing the largest difference according to the LFS figures are also countries with high shares of 'casual jobs and other arrangements'. If we leave out this flex-job type, we find that the ECHP figures are generally lower than the LFS figures. Therefore, it appears that the LFS includes some, but certainly not all, of the jobs being included in the 'casual jobs and other arrangements' category within the ECHP (see also Meulders et al., 1994; Bosch, 1995; Delsen, 1995). Insecure' Jobs The number of workers in non-standard employment is one indicator of a loose attachment to the labour market but a rather static one. Longitudinal measures of flexibility or insecurity are preferable but here the data limit the sort of analyses that can be conducted. Because of a lack of information on the type of contract in the first wave of the ECHP, it only appeared possible to look at the changes across the second and third waves. In Table 2 the transitions between the various working statuses between 1995 and 1996 are presented. 6 The LFS figures include all types of temporary jobs like fixed term contracts, temporary agency work, casual work and on-call contracts. 7 The term casual work has been used in the ECHP questionnaire to refer to contracts according to which people are hired by the hour of the day, with no promise about tomorrow or the next week. In this sense 'casual work' is different from a 'short-term' contract having a fixed term of one year or less. Schmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 Working Profiles and Employment Regimes in Europe 99 ploy merit record compared to the other countries because they have less secure employment but also less insecure employment and more people fully excluded from the labour market. From other sources it is known that the Southern states are characterised by strong employment regulations (OECD, 1999), less active labour market policies and high unemployment rates due to lower economic growth rates. The number of people in employment is generally lower, as is employment growth. This might point to a different type of employment regime. Subsistence security is not attained in Southern welfare states by a generous welfare system as in the social democratic regime, or in a highly flexible, efficiently operating labour market with low unemployment rates, as in the liberal regime, but by a highly regulated labour market with employment security. Employment security seems to be attained at the expense of a less efficiently operating labour market, because of which such a regime has to accept high levels of unemployment and inactivity. The picture emerging from the welfare state type of classification is that the Southern employment regime appears to be quite distinct from the continental corporatist and social-democratic regimes in the North. However, the liberal regime type is hardly distinguishable from the continental corporatist one. The liberal type certainly has less employment regulation but the share of temporary employment is substantial and larger than one might expect in a liberal regime where there is no need to attain flexibility through temporary jobs, since flexibility is innately achieved by a low level of employment protection and prevailing firm practices with respect to layoffs and quits in situations of demand cuts. The evidence found here, that the number of people with a weak attachment to the labour market appears lower in the Southern countries, does not mean that the labour market performs better. People need not stay unemployed for shorter periods in these regimes nor do they have more chances for escaping from precarious employment. Due to the lower economic and employment growth in the Southern countries the chances of moving upwards on the job ladder into permanent jobs is likely to be smaller than in the other European countries. To examine this issue further, the transition probabilities of moving upwards or downwards on the job ladder across the various countries are analysed. To what extent the employment regime is capable of guaranteeing that people move from partial or insecure employment into secure employment should be an important indicator for the labour market performance of welfare states. The upshot for these employment regimes is to what extent they permit people in partial employment or partial exclusion to escape from these precarious jobs and to move upwards into better, more secure jobs. Schmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 100 Ruud J. A. Muffels and Didier J. A. G. Fouarge The transitions across the three years by regime cluster are depicted in Table 5. In liberal, social-democratic and corporatist states people in full, secure employment in 1993 have slightly higher chances of remaining in stable jobs in 1994 than they have in Southern Europe. The findings show that in the South more people in full or secure employment are likely to move into full exclusion from the labour market the following year (10%, as opposed to 5-6%). Furthermore, partially employed and partially excluded people have much higher chances of escaping from unstable and moving into stable jobs in the liberal, corporatist and social-democratic states than people have in the Southern region. The Southern people with partial or insecure labour records in 1993 have much lower probabilities of getting out of that situation and moving into secure jobs. About 42% of the people in insecure jobs in the South are capable of moving upwards on the job ladder into stable jobs against 60% in the liberal states, 52% in the social-democratic countries and 46% in the corporatist countries (the European average is 50%). Further, the proportion of people moving from partial employment into full exclusion is still higher in Southern countries than in the other regions (34% in the South against 21% in the social-democratic countries and 28% in the liberal and corporatist countries). The conclusion must be that upward mobility is lower in Southern Europe and downward mobility higher. From a review of the evidence on the labour market performance of these employment regimes, it might be concluded that the Southern regime is performing worse in terms of enhancing job mobility and preventing labour market exclusion. What the reasons for these differences are, apart from differences in employment protection legislation, is left for further scrutiny. In the next section, this issue is examined by estimating explanatory models for explaining why some people are more likely to be work secure whereas others are more likely to be excluded from the labour market. Two types of models were estimated, (1) a model which explores the probability of belonging to any employment status (as defined on the 36 months of observation) and (2) a transition model exploring the changes in 12-month employment status between the first and last year. Determinants of Longitudinal Employment Status First we model the probability of belonging to one of the longitudinal employment status for the 36-month period. Results are given in Table 6. We contend that having a particular status is likely to be not the result of a random walk but affected by what happened before the observation date. We therefore posit the existence of a selection effect for which we need some Schmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 Working Profiles and Employment Regimes in Europe 101 Table 6 Estimation results of the multinomial model for explaining the belonging to the longitudinal employment status groups, 1993-1995 (36 months), Beta coefficients [t-values] (reference group: partial exclusion) Fully employed / Partially employed / Fully Excluded work secure work insecure Male Ref Ref Ref Female -1.284 [12.63]*** -0.552 [9.30]*** 0.903 [7.99]*** Age 0.502 [32.86]*** 0.139 [5.62]*** -0.043 [1.37] Age squared -0.006 [31.40]*** -0.002 [6.61]*** 0.001 [3.08]*** Low education level -0.219 [2.65]*** -0.063 [0.84] 0.059 [1.12] Average education level Ref Ref Ref High education level 0.396 [3.74]*** 0.418 [3.93]*** -0.672 [11.87]*** Number of children -0.095 [2.16]** -0.037 [1.02] 0.136 [2.47]** Household size -0.169 [4.40]*** -0.097 [3.70]*** 0.053 [2.05]** Single Ref Ref Ref Couple no child -0.181 [0.75] -0.069 [0.36] -0.072 [0.59] Couple with child(ren) -0.502 [4.60]*** -0.351 [2.25]** 0.092 [0.57] Lone parent -0.258 [1.28] -0.458 [3.95]*** -0.105 [0.81] Other -0.224 [1.37] 0.088 [0.56] -0.273 [1.75]* Never married Ref Ref Ref Married 0.467 [4.27]*** 0.323 [3.10]*** -0.281 [2.00]** Separated 0.672 [6.45]*** 0.382 [5.50]*** -0.351 [7.84]*** Widow(er) 1.195 [3.59]*** 0.551 [3.72]*** -0.632 [3.71]*** Life satisfaction 0.066 [5.99]*** 0.022 [3.00]*** 0.033 [2.09]** Bad health -1.054 [5.88]*** -0.328 [2.82]*** 0.291 [2.74]*** Capital income /1000 -0.024 [1.36] -0.031 [2.07]** -0.036 [1.52] Social transfer income /1000 -0.390 [7.98]*** -0.127 [8.37]*** 0.081 [5.53]*** Unemployment in past 5 years -1.733 [9.61]*** 0.028 [0.32] -1.142 [8.56]*** Home owner 0.040 [0.41] 0.007 [0.09] -0.050 [0.61] Liberal Ref Ref Ref Social democratic -0.332 [1.91]* -0.132 [1.74]* -0.235 [1.74]* Corporatist -0.046 [0.81] 0.029 [0.86] -0.049 [0.36] Southern -0.459 [3.34]*** -0.567 [5.50]*** 0.778 [4.30]*** Constant -6.500 [16.10]*** -1.424 [3.04]*** -0.196 [0.27] N = 77,856; Pseudo R2 = 0.2682) *** significant at 1%; ** significant at 5%; * significant at 10% t-statistics and significance levels based on robust estimate of variance accounting for clustering of observations within countries 1) Estimated model excludes Austria and Finland 2) The same model including country dummies instead of the welfare regime type generates a Pseudo R2 equal to 0.274 Source: Eurostat, ECHP 1994-1996, own calculations. Schmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 102 Ruud J. A. Muffels and Didier J. A. G. Fouarge correction procedure. The variable previous unemployment experience in the past five years has therefore been added to the model and we contend that it will affect the likelihood of belonging to the work insecure. This also, at least partly, allows us to control for business cycle differences across countries. The people having an unstable work history in the 36 months prior to the interview in 1996 will be more likely to occupy an insecure job in 1996 than people with a stable work history. We also included a life satisfaction score, since life satisfaction is expected to be at least partly the result of the current living situation and partly the result of what happened in the past. The variable is meant to capture part of the selection effect, though we could presume that the satisfaction score is a personal trait and that holding a more positive attitude towards life provides better chances in the labour market. The inclusion of age and age squared should measure the labour market opportunities by age as well as the declining labour force participation at higher ages. The human capital variables are included to take account of the differences in labour market opportunities by education level (Becker, 1964; Mincer, 1993). Three dummies were created for education level. The dummy for high education refers to people having finished higher vocational or university training. Low educated people are people who have not finished education beyond primary school. The reference group consists of people having finished all education levels in between (average education level). Further, we contend that these human capital variables play a role in the outcomes of the search process for a better job. Our conjecture is that people not satisfied with their current job due to being in a temporary or insecure work contract or being lowly paid, search for a better job. Better educated people are more likely to be successful in getting their preferred job. This hypothesis follows the well-known 'job search' theoretical approach. In the first generation models of job search only unemployment to work transitions were considered but in more recent approaches also on the job-search and job-to-job transitions are included (e.g. Lipmann and McCall, 1986; Mortensen, 1986; Narendranathan and Nickell, 1986; Burdett and Mortensen, 1998). The variables level of capital and social transfer income are incorporated in the model to capture the 'incentives-to-work' effect. A number of studies have shown that homeowners display a lower degree of mobility on the labour market (Narendranathan and Stewardt, 1996; Oswald, 1996). For this reason, we include a dummy for home ownership. A dummy for persons reporting (very) bad health is also included because it is believed that their health problems might impede labour market participation. We include clustered country variables to account for the effect of differences in employment policies or employment regimes. The variables are measured at the beginning of the observation period (wave 1). Schmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 Working Profiles and Employment Regimes in Europe 103 People with higher levels of education and those who are married have higher chances of being fully or partially employed and lower chances of being fully excluded. The odds of being partially or fully employed rather than partially excluded decrease when the person's household has more children or when the household lives in a country with a social democratic or Southern employment regime. In Southern regimes the likelihood of being fully or partially employed is significantly lower than in Liberal regimes. On the other hand we find that people living in a Southern regime compared to people in a Liberal regime face a much higher relative risk of being fully excluded. Both findings strongly suggest the Southern regime to perform worse in keeping employment high compared to liberal regimes. Women face a greater risk than men of being fully excluded, while older workers and people living in countries with a social democratic employment regime are less likely to belong to the fully excluded compared to people living in a Liberal regime. A better life satisfaction score favours being in secure employment. The effect of having social security income is ambiguous as it significantly reduces the probability of being work secure (disincentive effect) but it also decreases the probability of full exclusion. Transitions in Longitudinal Employment Status The use of the monthly calendar information permits us to look at changes in the longitudinal employment status between the first 12 months of observation at wave 1 and the last 12 months at wave 3. In modelling the transitions from one longitudinal status into another it is assumed that the explanatory model can by and large use the same sort of variables as in the previous model. Except, we now have the possibility of introducing some measures for the changes in household formation across the years. The multinomial model is aimed at explaining the transitions in employment status. If the status variable consists of four categories (secure, insecure, partially excluded and fully excluded), 16 types of transitions can be distinguished. Since our interest goes especially to transitions from partial exclusion to any other status, the set of statuses to be considered can be limited to three. The model9 has, thus, been estimated for three sorts of transition: (1) a transition from partial exclusion into secure employment, (2) a transition from partial exclusion into insecure employment and thirdly a transition from partial exclusion into full exclusion. The results of estimation are given in Table 7. 9 This time we estimated a multinomial model for the likelihood of making a transition between wave 1 and wave 3 (instead of belonging to a certain category as in the previous cross-sectional model). Schmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 104 Ruud J. A. Muffels and Didier J. A. G. Fouarge Table 7 Multinomial model for transitions out of partial employment, Beta coefficients [t-values] (reference group: stayers PE—»PE)1* PE—>FE PE —>PEX PE —>FEX Male Reference Reference Reference Female 0.058 [0.20] 0.239 [0.83] 0.558 [3.09]*** Age 0.011 [0.15] -0.050 [0.49] -0.239 [2.75]*** Age squared -0.000 [0.48] 0.000 [0.35] 0.004 [3.05]*** Low education level -0.070 [0.22] -0.062 [0.34] -0.198 [1.81]* Average education level Reference Reference Reference High education level 0.813 [2.37]** 0.425 [1.53] -0.238 [0.71] Number of children -0.231 [1.29] -0.215 [0.99] -0.069 [0.48] Household size -0.117 [1.63] -0.010 [0.11] 0.028 [0.31] Single Reference Reference Reference Couple no child 0.600 [0.65] 0.804 [1.13] 0.293 [0.56] Couple with child(ren) 0.282 [0.57] 0.466 [0.99] 0.293 [0.56] Lone parent 0.143 [0.28] 1.007 [1.38] 0.981 [1.49] Other 0.839 [1.72]* 1.005 [1.82]* -0.299 [0.47] Never married Reference Reference Reference Married 0.590 [1.60] 0.381 [0.99] 0.045 [0.18] Separated 0.495 [1.60] -0.182 [0.26] 0.136 [0.26] Widow(er) 2.818 [2.34]** 0.955 [0.65] 1.160 [0.80] Life satisfaction 0.025 [0.97] 0.019 [0.43] -0.010 [0.31] Bad health -0.221 [0.83] -0.027 [0.13] 0.782 [5.22]*** Capital income /1000 -0.189 [1.12] -0.285 [0.75] 0.037 [0.35] Social transfer income /1000 -0.036 [1.44] 0.006 [0.32] 0.006 [0.22] Job search -1.032 [3.21]*** 0.002 [0.01] -0.163 [0.88] Unemployment in past 5 years 0.183 [1.14] 0.408 [1.90]* -0.173 [0.84] Home owner -0.320 [2.51]** -0.571 [2.53]** -0.180 [1.10] Liberal Reference Reference Reference Social democratic -0.829 [4.30]*** -0.623 [3.11]*** -0.620 [2.22]** Corporatist -0.154 [0.41] -0.262 [2.07]** -0.074 [0.22] Southern -1.315 [4.21]*** -1.134 [3.40]*** -0.218 [1.00] Changes from wavel to wave3 No change Reference Reference Reference Married 0.519 [1.03] 0.090 [0.21] 0.293 [0.57] Separated -0.079 [0.10] 0.727 [1.17] 1.089 [1.98]** More children -0.168 [0.44] -0.163 [0.47] 0.107 [0.34] Less children -0.542 [1.52] -0.712 [1.24] -0.642 [1.18] More adults 0.575 [1.92]* 0.788 [1.67]* 0.321 [0.72] Less adults -0.070 [0.45] -0.088 [0.42] -0.208 [0.71] Constant 1.808 [1.44] 0.784 [0.49] 4.637 [3.18]*** N = 2,928, Pseudo R2 = 0.132 FE=full employment; PE=partial employment; PEX=partial exclusion; FEX=full exclusion *** significant at 1%; ** significant at 5%; * significant at 10% t-statistics and significance levels based on robust estimate of variance accounting for clustering of observations within countries 1) Estimated model excludes Austria and Finland Source: Eurostat, ECHP 1994,1996, own calculations. Schmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 Working Profiles and Employment Regimes in Europe 105 The models for explaining change generally have a lower fit than the cross-sectional models. There are various reasons for this: technical ones such as a low number of transitions, measurement error and attrition but also more substantial ones such as lack of information on the time varying variables that might explain the transitions from one longitudinal employment status into another. For the explanatory variables, information is available at the start of the observation period but not at the start of the transition. Nevertheless, the model presented here shows a reasonable good fit. The reference group is the group of persons not moving and staying in marginal employment or partial exclusion. The same kind of variables appear important for explaining the transition from partial exclusion into secure employment as in the model for partial exclusion (Table 6). Age has a positive effect and age squared a negative effect.10 A high education level has the expected positive sign. A lower education level diminishes the chances of making a transition into a secure job although the effect is insignificant. The same holds for the number of children. The variable, job search, has -contrary to our expectationsa negative sign indicating that the more one searches, the less likely he/she is to experience a transition into a secure job. Further inquiry into the reasons for this is needed. It might be that people in insecure jobs are not searching for a permanent job but prefer a non-standard job allowing them to combine working and caring duties, which is less feasible in a permanent job, because of long working hours. The employment regime variables appear to have a strong effect on the likelihood of a transition. The transition probabilities are somewhat lower in the social-democratic regime but markedly lower in the Southern regimes compared to the liberal regime. Hence, there is less mobility across employment status in the Southern regimes, corroborating our earlier results. The results for the other types of transition are more or less similar except that for the transition into full exclusion the social-democratic regime seems to perform best compared to the liberal regime. But even in the Southern regime the likelihood of moving into full exclusion is significantly lower than in the Liberal regime type. For the transition from partial employment into insecure employment, the unemployment history and the increase in the number of adults in the household during the three years exert a positive significant effect. Home ownership appears to have a strong negative effect on the transition into insecure employment. This corroborates the findings of others indicating that home ownership lowers the mobility on the labour market. 10 The point at the age curve where the positive effect of age reaches its maximum is 31 years for transitions from partial exclusion to full-time work and 37 years for transitions from partial exclusion into insecure employment. Schmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 106 Ruud J. A. Muffels and Didier J. A. G. Fouarge For the transition from partial employment into full exclusion the signs of the personal variables are likely to be the reverse of the signs for the transition into full employment. But for education level, separation and job search the signs are in an equal direction. The lower educated seem to have lower chances of moving into full exclusion than those with average education level. People searching more, have lower chances to move from marginal employment into full exclusion but the effect is insignificant. The findings for the clustered country variables show that the chances of moving out of the labour force into full exclusion are, as we expected beforehand, lower for the social-democratic regime compared to the Liberal regime. The dummy for people living in the Southern regime appears insignificant. 7. Conclusions and Discussion Existing cross-sectional statistics showed clearly that the labour markets across the various countries are rather different in terms of unemployment levels, extent of regulation, share of flexible labour, employment growth and share of the informal economy. Therefore, our purpose was to examine whether the use of longitudinal information would change the comparative picture substantially. Because of the short length of the observation period, particularly in the smaller countries, quite low numbers of transitions in terms of longitudinal employment status were observed. Partly due to this and theoretical reasons explained before, it was decided to cluster the countries into the four 'ideal-typical' classification derived from Esping-Andersen and others. Looking at the figures on flexible labour across the various countries a first test was obtained about whether the Southern countries should be considered a separate welfare state or employment regime or should be considered part of the corporatist type. It appeared that the share of flexible labour is indeed very much higher in the Southern countries. Next, looking at the distribution of employment profiles across the regime types, it was found that mobility from partial exclusion and insecure jobs into secure and permanent jobs is higher in the liberal and social-democratic countries than in the South. On the other hand, permanent employment appears to be more stable in Southern regimes. The opposite holds for flexible jobs being more fragile in the South. Both results might be due to the high level of employment protection regulation in the Southern region. The employment opportunities in the South to escape unemployment are generally fewer. Overall, remarkable stability with respect to permanent employment was observed. More than 85% of the permanent workers remained in their job between 1993 and 1995. On the other hand, there is substantial mobility beSchmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 Working Profiles and Employment Regimes in Europe 107 tween secure and insecure jobs. The image of a segmented labour market with secure jobs on the one side and flexible, insecure jobs on the other is far from reality in either employment regime. The closer the attachment to the labour market, the higher the income is for the various status groups compared to the people in secure employment. Next, looking at transitions from one longitudinal employment profile into another, confirms our previous results. Upward mobility is higher in the liberal and social-democratic countries and lower in the South. In addition, downward mobility is higher in the South. The Southern regime is, therefore, performing worse in terms of enhancing job mobility and preventing labour market exclusion. Although we only had three waves for analysing transitions across the various employment regime types, the results so far reported in this paper show remarkable differences across the three distinct 'ideal-types' of employment regimes. The analyses certainly need longer time-series to arrive at conclusions that are more robust, but the results are sufficiently encouraging to continue with this dynamic approach of testing regime-type differences using these excellent panel surveys. References Arts, W. / Gelissen, J. (1999). Three worlds of welfare capitalism or more? In search of ideal-types and real forms. Working Paper, WORC/Tilburg University, 21 p. Becker, G. (1964). Human capital: a theoretical and empirical analysis, with special reference to education. New York: National Bureau of Economic Research. Bonoli, G. (1997). 'Classifying welfare states: a two dimension approach. Journal of Social Policy, vol. 26, no. 3, pp. 351-372. Bosch, G. (1995). Flexibility and work organisation: follow up to White Paper Growth, Competitiveness, Employment. Report of Expert Working Group. Luxembourg: Office for Official Publications of the European Communities. Burdett, K. /Mortensen, D. (1998). Wage differentials, employer size, and unemployment, International Economic Review, vol. 39, no 2, pp. 257-273. Delsen, Lei, (1995). Atypical employment: an international perspective - Causes, consequences and policy. Groningen: Wolters Noordhoff. Esping-Andersen, G. (1990). The three worlds of welfare capitalism. Oxford: Polity Press - (1996). Welfare states in transition: national adaptations in global economies. London: Sage - (1999). Social foundations of post-industrial economies. Oxford: Oxford University Press. European Commission (1999). Employment in Europe 1998. Brussels: European Commission, Directorate V/A. Schmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54 108 Ruud J. A. Muffels and Didier J. A. G. Fouarge Ferrera, M. (1996). 'The 'Southern Model' of welfare in social Europe'. Journal of European Social Policy, vol. 6, no. 1, pp. 17-37. Goodin, B./Headey, B./Muffels, R./Dirven, H. J. (1999). Three worlds of welfare capitalism in panel perspective. Cambridge: Cambridge University Press. Leibfried, S. (1992). 'Towards a European welfare state? On integrating poverty regimes into the European Community. In Social policy in a changing Europe (eds. Ferge, Z. and Kolberg, J.). Frankfurt am Main: Campus Verlag. Lipmann, S. A JMcCall, J. J. (1976). 'The economics of job search: A survey'. Economic Inquiry, vol. 14, pp. 155-189. Meulders, D./ Plasman, O. & Plasman, R. (1994). Atypical employment in the EC. Aldershot: Dartmouth Publishing Company Mincer, J. (1993). Studies in Human Capital. Aldershot: Elgar Publishers. Mortens en, D. T. (1986). 'Job search and labor market analysis', In Handbook of labor economics vol. II (eds. Aschenfelter, O. and Layard, R.). Amsterdam: Elsevier Press. Muffels, R. (2000). 'Moving ages and edges. A lifecycle perspective on ageing and flexibilisation'. European Journal, London, vol. 6, pp. 32. Muffels, R./ Steijn, B. (1999). Flexible and permanent jobs on the Dutch labour market, empirical analyses of labour market flows and employment statuses using labour force surveys and panel data. AWSB Working Paper Series, 32 p. Narendrathan, W./Nickell, S. (1986). Estimating the parameters of interest in a job search model'. In, Unemployment, search and labour supply (eds. Blundell, R. and Walker, I.). Cambridge: Cambridge University Press. Narendranathan, W. / Stewart, M. (1993). Modelling the probability of leaving unemployment. Competing risk models with flexible baseline-hazards, Journal of the Royal Statistical Society, Series C, Applied Statistics, 42 (1), pp. 63-83. OECD (1999). Employment Outlook 1999. Paris: OECD Oswald, A. J. (1996). A conjecture on the explanation for high unemployment in the industrialised nations. Part I. University of Warwick Economic Research papers, no. 475. Verma, V./Russell, H./Nolan, B./Muffels, R./Fouarge, D./Fisher, K./Gershuny, J. (1999). Flexible labour and its impact on earnings and poverty. Evidence from wave 1 and wave 2 of the ECHP. Institute for Social and Economic Research / Colchester: University of Essex. Schmollers Jahrbuch 122 (2002) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.122.1.85 | Generated on 2023-04-04 12:29:54