Public Sector Employment in Europe: Where Will the Decline End?*

Franz Rothenbacher

The developments in the public sector have attracted growing attention in the last two decades. This interest is caused by the expansion of public expenditure, the growing state debt and the consequences of demographic aging and rising unemployment for the financing of social security and social services. Since the late 1970s governments have felt the need to control public expenditure in order to consolidate state finances. During the 1980s some success with respect to stabilizing or even reducing government outlays was achieved, although since 1989 public expenditure has been rising again.

1 Introduction

Public expenditure is closely related to public employment. Roughly 40 per cent of all public expenditure is spent in the form of salaries for public employees. If pensions are added, the share is even much higher. Thus one of the most important measures for curbing public expenditure is to halt or reduce public employment. Many governments have therefore tried since the early 80s to consolidate state finances by introducing measures to reduce public employment or to increase productivity in the public sector.

The close relationship between public expenditure and government employment can be seen in Figure 1.

Partially there is a relationship, but not entirely, as the rather high unexplained variance shows. There are congruent cases where state expenditure and government employment are high resp. the opposite, and there are also incongruent cases, where state expenditure is high, but government employment is low (the Netherlands, Italy); or, on the other hand, cases where state expenditure is low and government employment is high (Australia). Thus different models of state activity and, consequently, of public employment exist in industrialized countries.

Furthermore, the individual countries are at different levels of economic development, and therefore face different problems. Consequently, there are no universal solutions for all countries. Thus, the pattern of public employment in different countries remains complex. As a consequence, developmental trends have to be studied in relation to national variations. Therefore the attempt is made in this article to present a general framework by looking at international trends in government employment.

The main development in public sector employment can be characterized as a decline after a long growth period. This contraction is accompanied by a feminization of and an increase in part-time work in the public sector. Other developments are: an increase in the number of higher qualified employees, a decline in the number of non-permanent public employees and - most importantly - shifts in the functional composition of public employment.

2 Main trends since the mid-70s

Here the main empricial trends in advanced industrialized countries will be described and analyzed. First, the aggregate trend of public employment will be outlined. It will also be disaggregated by sex. Further steps will be a look at the relation of part-time to full-time work. Other important topics dealt with will be trends in different status groups in the public sector, especially the effects of education.

Diachronous developments: Growth, consolidation, devolution

The following section presents data on general government employment in the OECD countries and its historical development since the 1960s. It has to be noted that these employment figures refer to "general government employment", as defined by the OECD National Accounts statistics which corresponds to the UN National Accounts statistics. Therefore public enterprises producing and selling goods for the market are normally not included. Thus, these figures are not figures for "public sector employment" and are generally much smaller than public sector employment ratios, as for instance used by Rose et al. (1985).¹

After World War II the share of general government employment in nearly all developed industrialized countries was around 10 per cent. In 1960, of all European countries only Belgium and Sweden had shares higher than 10 per cent. On average, the share of general government employment nearly doubled from 11.2 per cent to 18.2 per cent in all 15 EU countries. The growth rates between the individual countries were very different in this respect. The highest rate could be found in the Nordic welfare states. In Denmark the share doubled, in Finland the share trebled, and in Sweden the share almost trebled from a very high level. All countries with a share of 30 per cent and more belong to the group of Nordic countries. On the other side of the spectrum - the countries with the lowest employment shares - we find Switzerland (12.0 per cent), the Netherlands (12.8 per cent) and Luxembourg (10.9 per cent) within the group of developed industrialized countries. These countries have, in addition, other common traits. They belong to the group of countries which are known for their low "stateness", id est, the central government has rather few competences, and liberalism has a strong position (Badie/Birnbaum 1983; Heper 1987). Another group of countries with a lower degree of economic development - the European peripheries, namely Ireland (14.0 per cent), Portugal (13.6 per cent), and Spain (15.1 per cent) have remarkably smaller employment ratios in general government employment. A historical description of the direction of trends and differences in the levels of the overall employment in general government shows that in all OECD countries the overall employment in the post-World War II period went up until it reached an upper limit and meanwhile goes down or stagnates. At the end of the 1980s, following the example of the United Kingdom, several countries started to reduce public employment. From 1989 onwards the growth rates were negative in the following countries: the United Kingdom -6.1, Denmark -0.4, Finland -1.2, Sweden -2.1, the Netherlands -0.8. A zero growth rate could be found in Italy and the Federal Republic of Germany. Thus, the employment figures were reduced especially in those countries that have the highest levels of general government employment - with the exception of the United Kingdom.

The developments in Eastern and Southeastern European transition countries differ from developments in West European countries. While in Western Europe a trend towards stagnation predominates, a strong increase in public employment can be observed in Eastern Europe. In Romania, for instance, public employment rose from 724,270 in 1988 to 899,022 in 1991, a rise of 24 per cent (ILO 1993: 12).

As OECD figures based on national accounts do not show the extent of total public sector employment, one should use an alternative source for obtaining public employment figures. On the level of the European Union the EU Labour Force Survey (LFS) is a highly useful instrument for employment questions. But the LFS has the disadvantage not to supply appropriate public sector employment figures, at least not from the printed sources.

There are two possibilities: (1) One way to get figures is to quantify public sector employment in a definition which includes the public sector as a whole, that means, it also includes private sector social services activities. In principle, the data allow a description of the levels of the individual countries, of global trends and convergence resp. divergence processes. The pattern does not match in every respect the one received by the OECD general government figures. Here, Denmark ranks highest as a Scandinavian country, the Netherlands take the second place, which is surprising, because general government employment is low in the Netherlands. One explanation could be that private services are much more important than public services in the Netherlands. Belgium's figures are higher than France's in this respect, and the United Kingdom's lower than France's. The Netherlands and Italy appear once more in the midfield. Ireland and the Mediterranean countries Portugal, Spain and Greece have the lowest shares (in descending order). In relative terms, the individual countries have shares ranging from about 15 per cent to 35 per cent. In diachronous development there seems to be convergence to some extent, as the most developed countries have reached their upper limits and partly have declining shares, whereas the less developed countries, especially Greece, have overproportionate growth rates. In the Mediterranean countries and in Ireland the large, but declining primary sector still exerts its influence on these figures.

(2) Another possibility is to take the data for public administration in a narrow sense. If one therefore looks at employment in public administration only on the basis of the EU Labour Force Survey, a different ranking of countries emerges. Here Denmark, the United Kingdom and the Netherlands have very low proportions. As these figures do not include public and private social and other services, one could assume that these services are very important in Denmark, the United Kingdom and the Netherlands, and that public administration (including defence and compulsory social security) is generally less important. Most of the other countries are rather congruent cases, with Belgium and France having high figures in administration and public services, and the Mediterranean countries and Ireland again having low figures according to both calculation methods. Figure 2 presents public sector employment in the five European countries of Belgium, France, Germany, Sweden and the United Kingdom. The data used are national data and therefore subject to national definitions. With the exception of Germany, the data refer grosso modo to public sector employment, while for Germany only data on the public service are available. As we can see, public employment, starting from different levels, expanded in all these European countries until the 1980s. Governments tried to cut back public employment - also to varying degrees since the late 1980s.

Structure of employment in the public sector

After having looked into the development of public employment over time with different indicators, I will now analyze the structure of public employment. In a comparative attempt, unfortunately, it is only possible to compare some few structural features of the administrative systems. The structure of public employment can be defined as the division of the publicly employed by sex, duration of work, professional status and career group. Comparisons can only be made here with respect to sex and duration of work. A comparison of the functional division of public employment, as done by Richard Rose et al. (1985) is not yet possible with international data. But female employment and part-time employment are two important indicators for developments in public employment and shed some light on changes in government functions, especially the growth of social services in education and health. Thus, the following sections will describe the feminization of public employment and the increase in part-time work in public employment.

Feminization: Growing share of women in public sector employment

Growing female work participation in the whole economy is a feature common to all European countries. These developments are well-documented by the annual reports of the "Employment Observatory" of the European Union (European Commission 1994). In most European countries the overall activity rate of women ranges from 40 per cent to 50 per cent, but it does not exceed 50 per cent in any country. As in the economy as a whole the share of women in public sector employment is constantly growing throughout Europe. Government and public services have come to be the most important employers for women in the last decades.

The differences regarding the overall female employment share in the public sector are tremendous. In many Western European countries, the absolute employment figures in the public sector are meanwhile going down in favour of absolute employment increases for women. This is the case with the Netherlands, Sweden, New Zealand, and more or less the US (ILO 1993: 12).

The question arises, how many (in relative terms) of all women in employment work in public administration and other services. This indicator also informs us about welfare state development or the development of social services in general. According to this figure, the importance of government and social service employment varies to a high degree within the European Union and has reached the upper limit in many countries, as the low growth rates show. In the case of Denmark and the Netherlands, over 50 per cent of all women employed work in these sectors. In Greece, this figure is extremely low, with only one quarter of all employed women being employed in those two sectors. France and Belgium have values of around 40 per cent. In most countries it varies between 35 and 40 per cent. The share is smaller in the Mediterranean countries and in Ireland. Another possibility for measuring female employment in the public sector is to recur to national statistics where available. It is important to note that national public sector statistics with a detailed socio-structural differentiation of categories are rather rare. Those countries for which data have been available are presented in Figure 3. This figure shows the proportion of women employed in the public sector according to national definitions and published results. The interpretation has to be given on the basis of national definitions. Developments over time are more reliable than differences between countries. Principally, the data show the importance of female public sector employment in total female employment. In Sweden half of all employed women have public sector jobs. After a steep rise in the 1970s and a slowdown in the early 1980s this high proportion was reached by the end of the 1980s, and it has been stagnating since then. In the other countries (France, Belgium, Germany, United Kingdom) an increase in female work participation occurred in the 1970s, and a stagnation in the 1980s. The only exception within this group of five countries is the United Kingdom where female work participation has declined since 1985. Given the higher overall employment rate of women in France compared with Belgium, it is interesting to see that public sector employment of women is obviously more important in Belgium than in France. In the United Kingdom female labour force participation in general is rather high and higher than in Belgium. Thus, the drop in public sector female employment below the Belgian level suggests that a shift from the public to the private sphere has occurred. In Germany, only roughly 15 per cent of all women in employment work in the public services, which is a narrow definition.

Figure 4 shows that if the proportion of women in public sector employment, based on national sources, is measured, it becomes clear that feminization is highest in Sweden with over 70 per cent in 1995, followed by the United Kingdom and France with roughly 60 per cent, and Belgium and Germany with roughly 45 per cent in the early 90s. The nearly linear trend of this indicator in every one of the five countries since the 60s is most impressive, while the levels of the individual countries remain stable. Thus, there is clear evidence of an ongoing feminization of the public sector. In most countries there are no signs - not even in Sweden - that an upper limit has been reached.

The pattern changes once more if one looks at public administration alone, according to the EU Labour Force Survey. Once again a change in the ranking order takes place. France and Belgium now have the highest shares, the Netherlands and the United Kingdom have rather small ones.

The alternative indicator (based on the EU LFS) measuring the feminization of employment in public administration and other services in terms of the proportion of women in these two sectors answers the question to what degree women predominate public sector employment. With over 70 per cent feminization is highest in Sweden and Finland; with nearly 70 per cent it is still high in Denmark and the United Kingdom. Surprisingly, Portugal also has a rather high proportion with 65 per cent, probably a consequence of its overall high female employment ratio. Most continental and Mediterranean countries have proportions around 55 per cent. The female employment rate is low in Italy, Luxembourg and Greece. But in nearly all EU countries women are now in the majority in these two sectors.

The feminization of public administration alone is much lower, partly because the male-dominated defence sector exerts its influence. Only in Denmark has this share reached the 50 per cent level. The development over time reveals that there is convergence to some degree, because some countries have reached their upper limits, while others are catching up a little bit. The countries with the highest feminization ratios are the Nordic countries, France and the United Kingdom. Again, the Netherlands have a very low rate in this respect.

Segregation: Male employment in public administration and other services

If one looks at the sexual division of public employment, one should not forget male employment. The tertiarization of the economy with a declining industrial workforce should lead to a growing number of men working in the service sector. Is this also true for the public sector? According to EU Labour Force Survey data, in the 1980s the shares in these two sectors increased also for men. Some results are surprising. In Belgium, the Netherlands and Italy nearly every fourth man works in these sectors, in Italy a steep increase occurred in the 1980s. Most continental countries and the United Kingdom are in the midfield with 15 per cent to 20 per cent of total male employment. The Mediterranean countries and Ireland have the smallest proportion of men in these sectors.

If one looks (according to the EU LFS) at public administration alone, only 5 to 12 per cent of all men work in this sector in EU countries. As defence is included in these figures, the variation is presumably heavily influenced by the size of the military forces. Belgium and Germany have the largest share in this respect. Denmark has one of the smallest shares. A declining trend can be observed in several countries (Germany, Belgium) and an increase in others (Greece, Spain).Whereas the public sector becomes increasingly important regarding the occupational opportunities of women, public sector employment is becoming less important for men. This can be seen in Figure 5 which is based on national sources. In all five countries the share of publicly employed men in total male employment increased until the mid-1980s and subsequently declined partly dramatically. In Sweden, high female employment in the public sector does not coincide with high male employment in the public sector. According to sector of employment job segregation seems to be high there. Male employment in the public sector was very high in the United Kingdom, but lost 10 percentage points. The Belgian male public sector employment figure is still very high, but strongly declining, too. The French male public sector employment figure is declining to a small degree, but the decline is stronger in Germany. Most impressive is the declining share of males in Sweden and the United Kingdom.

Replaced by what? Growth, stagnation and decline of public employment

Different developments in female and male public employment, stagnating employment for women, and declining employment for men, result in a overall decline of public employment. According to Figure 5 the developmental paths of all countries show one similar pattern: growth, consolidation and decline. Due to different traditions and structures, the levels of the individual countries are very different. The figure is very similar to the one known from diffusion theory or from the literature on growth processes in general, where one "product" experiences mostly a "logistic" growth process and afterwards a decline due to it being replaced by other products. In this case the product would be "public goods". The question is what could be the substitutes? The processes which are responsible for the substitution of private employment through public employment are processes labeled "privatization", "decentralization", "slim state", "efficiency", "competition", etc. The question which could be discussed is how far the public sector is able to retreat from the production of "public goods", and to what extent "public goods" can be produced on a private basis. The second subject for discussion would be to what extent "external" processes can be considered to be responsible for a reduction of public services. One of these processes is the end of the cold war, the opening and reshaping of frontiers, which reduces the necessity of defence. As a consequence of these developments the military personnel has been reduced nearly everywhere in Europe. The third topic for a discussion are the consequences of European integration. The ongoing integration process tends to weaken national responsibilities and to shift responsibilities to higher levels. It therefore reduces the demand for national and "nationalized" services as, e.g., the national postal services, the national railways, etc.

Increasing part-time work in public sector employment

• General development. As in the economy as a whole, part-time employment is also growing in the public sector.² Before we discuss the development of part-time employment in the public sector, we should look at part-time work in general. The increase in part-time work in the whole economy is a result of the increase in female labour force participation. The main group of part-time workers are women. The share of part-time working women in relation to all persons in employment is a first indicator of the importance of female part-time work. If the number of men working part-time is compared with the total number of men in employment it becomes clear that men working part-time are the exception.
• Variations. With only few exceptions part-time work is increasing in all European countries, but inter-country differences are rather big if part-time work is put in relation to total employment. For 1990 some country clusters can be distinguished: part-time work is high in all developed welfare states, that is in Scandinavia with the exception of Finland, where full-time work dominates. The highest extent of part-time work is to be found in the Netherlands. In the European peripheries, the Mediterranean countries and Ireland, part-time work is rare. Some continental countries, such as Germany and France, occupy a position in the middle (European Commission 1994; Ellingsæter 1992).
• Sex differences. Major differences can be found with respect to sex. Part-time work is predominantly a matter of female employment. As a rule, the share of part-time working women is high if female work participation is high, as in the Nordic countries. On the other hand, in countries where female labour force participation is low in general, part-time work is low, too. Again, the Netherlands are an exception, with the overall female employment ratio being rather low, but part-time employment being the highest in Europe. Part-time working men are a minority in all European countries, but the importance of part-time work for men is growing, too. It is most widespread in the Nordic countries, but the biggest share of part-time working men is to be found in the Netherlands. In the Nordic countries part-time work of men is more common. In the peripheries part-time work of men is mainly a result of latent unemployment and underemployment.
• Women’s share in part-time employment. The general feature of rising part-time employment but with different weight for the sexes can be supplemented by the question, how total part-time is composed. How many part-time jobs do women have as compared to men? Women in all European countries are dominant in part-time employment; it varies from roughly 70 per cent to 90 per cent. With about 90 per cent, the female share is highest in Belgium and the Federal Republic of Germany. Male participation in part-time employment is therefore, in contrast, lowest in these two countries.
• Part-time employment in the public sector. Empirical information on part-time employment in the public sector has only been available by means of national sources so far, mainly the national Labour Force Surveys and administrative statistics. In principle, part-time employment is included in the EU Labour Force Survey, but the published results do not include part-time employment in "public administration" and the "other services" sections. But even the national publications normally do not inlcude figures on part-time employment. Therefore data can only be supplied for some countries. Figure 6 shows all public sector part-time employees as a share of all public sector employees. In principle, since 1945 the overall trend has been a rise in part-time employment in the public sector. Growth rates of part-time public sector employment were higher than growth rates of the whole public sector. The decline in public sector employment in the last ten years has increased the importance of part-time employment. Thus, in the United Kingdom public sector employment has decreased strongly since the 80s, but part-time employment has gone up by more than 10 per cent. There are some interesting national differences. In France, part-time employment was rather high as early as the 1950s, as compared to Germany, for instance, where it was negligible. In Sweden, Germany and France the shares of part-time employment in the public sector are similar now (1990s).
• Sex differences. There are major differences regarding sex in part-time employment in the public sector. If one compares Figures 7 and 8 it becomes clear that in the public sector - as in the private sector as well - women represent the majority of part-time workers. Country differences concerning female part-time work are high. Every second woman in the public sector in the United Kingdom (1990s) works part-time compared to roughly 25 per cent in Sweden. In Germany nearly two in five women in the public service work part-time. Thus, there are different models of public sector employment for the individual countries. In Sweden, full-time public sector employment of women is predominant due to extended provisions for the reconciliation of family tasks with work. In the United Kingdom, such provisions are obviously available to a much smaller extent, and Germany is somewhere between these two countries.
  

• Part-time employment of men in the whole economy as well as in the public sector can be used as an indicator of parental leave arrangements, also allowing and stimulating men tor reconcile family and work; but it could also be a result of unemployment or the reduction of male labour supply in case of divorce. Figure 8 shows that Sweden is most advanced in this respect, with almost 12 per cent of all men in the public sector working part-time. In one year only (1993-94), due to the reduction of public sector employment, the share of part-time working men in Sweden rose by 2 per cent. In the United Kingdom, even less than 10 per cent of all men in the public sector are employed part-time, but the proportion is increasing strongly, probably because of a similar reduction of public employment. Germany maintains the traditional model of very low or non-existent male part-time work in public services in this respect.
• Distribution by sectors. Apart from aspects of growth and sex division, the distribution of part-time workplaces in different fields of the public sector is of interest. Here data are available for the United Kingdom and for Germany. In the United Kingdom, given the overall decline of public sector employment, part-time work has been rising tremendously in the last ten years (1985-1995) by almost 10 percentage points. Part-time workplaces have been reduced in central government and in the National Health Service. Especially public corporations as well as other sectors, e.g. education, have increased their part-time labour force. In Germany part-time work (part-time workers T1 in per cent of part-time workers T1 plus full-time working civil servants, judges and public employees in the direct public service; the figures are higher in this calculation than if all public service employees would be counted) is generally highest in the communes and the communal associations, whereas part-time work in federal government only amounted to 10 per cent in 1994. In the Federal Railways part-time work was in principle non-existent, but it was widespread in the Federal Post. If only women are considered we get the same structure. Part-time working women in federal government accounted for 30 per cent of all women employed in federal government. In the communes and the Federal Post part-time work of women is widespread. 40 to 50 per cent of all women work part-time there.

Relative decline in number of "non titulaires"

In most European countries public employees are grouped in different status categories. Civil servants in the sense of permanent civil servants ("fonctionnaires", "Berufsbeamte") are a minority in several countries, at least in the United Kingdom, where civil servants account for only 10 per cent of all public employees, and in Germany, where civil servants ("Beamte") have to be distinguished from employees ("Angestellte") and workers ("Arbeiter") in the public sector. The small nucleus of civil servants in Germany conducts sovereign tasks ("Hoheitsaufgaben"), is subject to public law and is permanently employed. The employees and workers in the public sector are not employed on a permanent basis (de jure) and are subject to private law. Besides these three main categories, in Germany there are also civil servants and public employees who are on probation or are employed temporarily or for specific tasks only. In France most public servants are employed on a permanent basis and subject to the civil service code, but there exists a group of non titulaires who are employed temporarily or for specific tasks.

A trend which can be found in several European countries is the relative reduction of "non titulaires". Titularisation means the transformation of temporary employment contracts into permanent civil servant positions. This procedure is common especially in France and Italy. In the French system only fonctionnaires, ouvriers d’état and praticiens hospitaliers statutaires exist. Non titulaires are those employees who do not fit in these groups, that is, especially persons who are employed for limited time periods or have part-time jobs; therefore they are not intended to be given the status of "fonctionnaires". In France the number of non titulaires did not change significantly between 1969 and 1989: 864,000 in 1969 and 877,000 in 1989. As the number of total employees in the public sector rose steeply, the relative share of non titulaires dropped from 28.4 per cent to 17.9 per cent in the same time period. In the French public sector the number of non titulaires varies greatly. In the hospital sector the non-medical staff normally does not belong to the non titulaires (5 per cent). Physicians practising at hospitals normally are non titulaires, because, as a rule, they have their own practices and are therefore in principle self-employed. There have been several plans in the ministries to reduce the number of non titulaires. In 1969 456,000 non titulaires were employed in the ministries, in 1982 the number was highest with 481,000 employees. By 1989 the number of non titulaires had been reduced to 386,000. The composition of non titulaires has changed fundamentally in this respect. The number of part-time employees in the group of non titulaires has grown overproportionately. Meanwhile there is one part-time non titulaire for each full-time non titulaire. In the field of collectivités locales the number of non titulaires is by far the highest. The recruitment of non titulaires increased in the 80s with ups and downs in absolute terms, accompanied by a relative decline due to the even stronger increase in positions for fonctionnaires. With 380,000 non titulaires by the end of 1989 the fonction publique territoriale had a share of employees outside the civil service code of over 30 per cent (Quarré/INSEE 1992: 46f).

From a comparative perspective it is difficult to find out where changes from temporary employment into civil service employment took place due to the heterogeneity of the systems. Processes of titularisation can only be observed in those administrative systems which are similar to the French one. In other systems as, e.g., the German system, there is a clear demarcation line between civil servants (Beamte), public employees (Angestellte) and public workers (Arbeiter). To change a position planned for a public employee into a civil servant position is only possible through political decisions. Titularisations become interesting with respect to the functioning of the political system when they are used to reward the clientel of the winner of the elections and to supply them with civil servant positions. Non titulaires can therefore be regarded as a potential group for the recruitment of civil servants. In France (Quarré/INSEE 1992: 47), Spain (Martin-Retortillo Baquer 1992) and Greece (Spanou 1992) civil servants have been recruited in the group of non titulaires.

Increase in number of highly qualified employees

Another trend is the permanent increase in qualification levels. Nearly every system of public administration is structured in such a way that a position in the administrative hierarchy is linked to the achieved educational level. The expansion of the educational system in general and the fast integration of women into the educational system leads to a rise in the number of higher educational degrees and thus to an oversupply of highly-qualified persons. This, in turn, affects the recruitment methods, i.e., an increasing number of positions are assigned to persons with higher educational degrees. In the final analysis this can be interpreted as a market imbalance resulting from the oversupply of candidates for only a small number of positions; thus, those with higher grades get the jobs. This in turn circularly affects the requirements for higher positions. Thus it can be said that ceteris paribus this has important consequences for pay and pensions.

• France. First, one has to distinguish between the qualifications required for positions in the public administration and the qualifications of the applicants. In France, for instance, in the annual "concours", it is increasingly applicants with educational degrees that are higher than those required for the positions in question who compete for the jobs. This could be a sign for an oversupply of highly qualified persons. It can furthermore be said that qualifications change significantly while the grade system lags behind reality. Thus the actual profile of the whole group might change due to the oversupply of highly qualified persons. The pattern of hierarchical categories becomes even more complicated if the titularisations are included in the lowest category (D). On the whole, since 1969 an increase in employment in the highest category A has taken place, whereas category B has not changed substantially, and categories C+D have clearly diminished, with category D nearly dying out. The strongest pushes in the qualification structure occurred after May 1969 and May 1982, the time points with the highest number of titularisations. Another aspect are the different qualification structures of the three parts of the public sector. In state employment, grade A and B account for 60 per cent of all employees (teachers, officers and noncommissioned officers, administrative staff); in hospitals for 40 per cent (physicians and medical staff), in collectivités locales for 20 per cent (Quarré/INSEE 1992: 48f; Rouban 1995).
• Germany. In Germany there are four career groups: the higher service (höherer Dienst), the upper service (gehobener Dienst), the middle service (mittlerer Dienst) and the lower service (einfacher Dienst). These career groups are identical for civil servants and public employees. Career groups are linked to educational degrees. The higher service is open only for applicants with university degrees, in the upper service the upper secondary school (Abitur) is required, in the middle service it is the secondary school; in the lower service primary school is the minimum prerequisite (Southern 1979: 137; Schmidt/Rose 1985).

In general, in Germany as well as in France, a relative growth of upper career groups can be observed, that is, the higher service and the upper service. This is the case with civil servants as well as with public employees. The higher service has nearly doubled its share since 1960, while the upper service has grown only little, while the middle and lower services have declined. The structure of career groups in the institutional parts of the public sector has remained the same. The higher service is strongly represented in state government, probably because of the teachers, who are employed by the states (Länder). The federal government (Bund) employs considerably fewer persons in the higher service, and the communes even fewer. It is important to look at the absolute numbers in this respect. In 1994, the federation employed 198,000 civil servants/judges and public employees, the states 1,400,000, the communes 726,000, the communal associations 30,000, the Federal Railways (Deutsche Bundesbahn) 121,000 and the Federal Post (Deutsche Bundespost) 304,000. The central government is therefore rather small and accounts only for one sixth of all state employees. Thus, the communes employ half of all state employees. The centre of gravitation of the German administrative system is therefore the state government where, in addition, the labour intensive fields of education and health are located.

3 Conclusion

1. The trends in public sector employment are not uniform. Some kind of consolidation or stagnation or even small declines have occurred in most countries. In only one European country - the United Kingdom - has a structural decline taken place. In other countries, especially Finland, Iceland, France and Austria an ongoing growth of public employment can be observed. In the last years, especially Switzerland has modernized and extended its public administration. There are signs that Sweden has reached the upper limit of government employment, which is around 30 per cent of employment in general government. Norway and Denmark have also reached this level meanwhile. The question is if other countries - besides the Nordic ones and possibly France - will ever reach such levels.
2. Where aggregate trends are not coherent throughout Europe, there is a common trend in public sector employment, namely increasing feminization. And there seems to be one regularity: the larger the public sector is on the whole, the higher is the female share. The underlying logic is probably the following. The traditional government functions require only a relatively small number of public employees. But as the welfare state expands and therefore social services as well, especially health and education, it is predominantly women who are recruited for these positions (Kolberg 1991; Pierre 1995). Thus, another rule can be derived: the lower government employment is, the more it is restricted to the classic functions of government and the higher is the share of male employees. The growth of the welfare state is congruent with the growth of female public sector employment. But there are exceptions, as the Dutch example illustrates, and the crucial variable for public sector growth seems to be the extension of the social services, which are labour intensive. Welfare states thus can be shaped in different forms, they can be mainly service oriented or place more emphasis on fiscal welfare.
3. Another coherent trend is the increase in part-time work. This phenomenon is strongly related to the growth of female public employment. Public employers give priority to part-time work. It functions as an employment policy and family policy measure for achieving a reconciliation of family obligations with female employment. Thus: the higher "welfare stateness" is, the higher is female public sector work participation and the larger is the share of part-time work (examples are the Netherlands and, partly, Switzerland).
4. One further common trend is the permanent increase in the number of higher qualified employees in the public sector. This phenomenon is not easy to explain, because different mechanisms can be at work: firstly the expansion of the educational system with a rising number of university graduates exerting pressure on public sector employment. On the pull-side there exist rising qualification requirements to fullfill a task. One important factor causing the shift in qualifications is the growing number of teachers who are mostly university graduates.
5. Public sector developments therefore have cross-cutting tendencies. Traditional government functions, such as defence, post and telephone, railways, either lose their traditional importance or have been privatized. On the other hand, due to the consequences of persistent low fertility, demographic aging and globalization, to name a few important trends, social services have been and will be expanded. This is especially the case in the fields of child care, old age care and health care. Thus public services and public sector employment are not being dismantled, it is rather their composition that changes.
6. A more general perspective of influences on national public administrations can be seen in the process of European integration. It can be supposed that the state formation process of the European Union has impacts on national administrations. The more the integration proceeds, pulling down internal boundaries and erecting external ones, the more the sovereign state functions developed by the 19^th century nation states are transferred to the European level or are subject to privatization. To be more abstract, one could also say that differentiation on a higher level requires dedifferentiation on a lower level. In this respect, the sovereign tasks of defence, post and railways are questioned. Especially in these fields decisive changes have taken place in nation states. One general trend is the reduction of the military staff, which is not only the result of the collapse of communism in Eastern Europe, but also of the tendential shift of the "national security" function to higher-level institutions. The relative as well as absolute reduction of the armed forces can be observed, for instance, in France and Germany. France has abolished the compulsory military service and changed it into a professional army. In Germany, the armed forces have been reduced substantially. Similar processes can be observed regarding the postal services and railways. In the 19^th century the nation states took over post and railways in order to integrate these areas economically and to institutionalize national defence. In a much larger economic and defence space the necessity to maintain these functions as national functions no longer exists. The privatization of some of these state functions is thus possible and has been initiated in several European countries.

Notes

• * A longer version of this contribution was presented to the ECPR Workshop on "Working for Government: Explaining Public Sector Employment Trends in the Last 20 Years", in Berne, Switzerland, 27th February to 4th March 1997.
• 1 According to the System of National Accounts of 1993 (Commission of the European Communities et al. 1993) the general government sector consists of four sub-sectors: (a) central government, (b) state government, (c) local government and (d) social security funds. The general government sector does not include public corporations, neither of the non-financial nor the financial sector (e.g. central banks).
• 2 There are usually two variants of the definitions and data on "part-time work": 1. Recalculation of part-time places in full-time places, or 2. each part-time workplace is counted as one unit and is classified according to the number of hours.

References

• Badie, Bertrand and Pierre Birnbaum 1983: The Sociology of the State. Chicago and London: The University of Chicago Press (originally published as "Sociologie de l’Etat", Paris 1979).
• Commission of the European Communities - EUROSTAT/IMF/OECD/UN/ World Bank 1993: System of National Accounts 1993. Prepared under the auspices of the Inter-Secretariat Working Group on National Accounts. Brussels/Luxembourg, New York, Paris, Washington, D.C.
• Ellingsæter, Anne Lise 1992: Part-time Work in European Welfare States. Denmark, Germany, Norway and the United Kingdom Compared. Oslo: Institute for Social Research (ISF).
• European Commission, Directorate General for Employment, Industrial Relations and Social Affairs 1994: Employment in Europe 1994. Luxembourg: Office for Official Publications of the European Communities.
• Heper, Metin (ed.) 1987: The State and Public Bureaucracies. A Comparative Perspective. New York/Westport, Connecticut/London: Greenwood Press.
• International Labour Organization, Sectoral Activities Programme (ILO) 1993: Joint Committee on the Public Service. Fifth Session, Geneva, 1994. Report I, General Report. Geneva: International Labour Office (JCPS/5/1994).
• Kolberg, Jon Eivind (ed.) 1991: The Welfare State as Employer. Armonk, New York: M.E. Sharpe.
• Martin-Retortillo Baquer, Lorenzo 1992: Pervivencias del "Spoil System" en la España actual. Anuario de Derecho Constitucional y Parlamentario 1992, no. 4, 31-59.
• Pierre, Jon 1995: Governing the welfare state: public administration, the state and society in Sweden. In: Jon Pierre (ed.), Bureaucracy in the Modern State. An Introduction to Comparative Public Administration. Gower House, Croft Road, Aldershot: Edward Elgar Publishing Ltd./ Brookfield, Vermount: Edward Elgar Publishing Company, 140-160.
• Quarré, Dominique and Institut National de la Statistique et des Etudes Economiques (INSEE) 1992: Annales statistiques de la fonction publique 1945-1969-1989. Paris: INSEE (=INSEE Résultats no. 169-170; Emploi-revenus no. 28-29).
• Rose, Richard, Edward Page, Richard Parry, B. Guy Peters and Andrea Gendali Pignatelli 1985: Public Employment in Western Nations. Cambridge: Cambridge University Press.
• Rouban, Luc 1995: Public administration at the crossroads: the end of French specifity? In: Jon Pierre (ed.), Bureaucracy in the Modern State. An Introduction to Comparative Public Administration. Gower House, Croft Road, Aldershot: Edward Elgar Publishing Ltd./Brookfield, Vermount: Edward Elgar Publishing Company, 39-63.
• Schmidt, Klaus-Dieter and Richard Rose 1985: Germany: the expansion of an active state. In: Richard Rose et al., Public Employment in Western Nations. Cambridge: Cambridge University Press, 126-162.
• Southern, David 1979: Germany. In: F. F. Ridley (ed.), Government and Administration in Western Europe. Oxford: Martin Robertson, 107-155.
• Spanou, Calliope 1992: Politique du recrutement dans l'administration hellénique. Une modernisation impossible? Revue Française d'Administration Publique no. 62, avril-juin 1992, 325-334.

Dr. Franz Rothenbacher
University of Mannheim,
Mannheim Centre for European
Social Research, EURODATA,
D-68131 Mannheim, Germany
Phone: +49-621-292-1738
Fax: +49-621-292-1723

E-mail: Franz.Rothenbacher@mzes.uni-mannheim.de

Franz Rothenbacher is a sociologist at the EURODATA Research Archive at the Mannheim Centre for European Social Research (MZES) and managing editor of this Newsletter