TEORIJA IN PRAKSA let. 56, 2/2019 660 Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC* THE CRISIS OF COLLECTIVE EMPLOYMENT RELATIONS SYSTEMS: FIVE COUNTRIES COMPARED Abstract. The present article looks at the situation of collective employment relations in Slovenia, Croatia, Serbia, Hungary and Austria in the period 2013–2015 following the crisis. A review of empirical data shows the countries in our sample differ greatly in the stages of (under)development of collective employment relations and/or dynamics of their destabilisation. The varying levels of both unemployment pressure and trade union power (or lack thereof) explain these differences. The analysis shows the mechanisms destabilising the collec- tive employment relations systems – which underpin such differences – are in essence similar across the sam- ple with characteristic similarities being found in the same sectors in the countries. Keywords: Collective employment relations, trade unions, works’ councils, numerical flexibility Introduction A chapter in the book Razpotja in prelomi (Crossroads and Shifts) states the core elements of collective employment relations (“collective ER”) are (relatively strong) trade unions and complementary institutions that enable employees to participate in decision-making processes on the micro level, and sectoral collective bargaining and social concertation on the macro level. Further, the chapter notes the “foundation (and the result) of such an institutional arrangement (…) is safe/stable employment for most of the active population”, and that the “prerequisite for that foundation (…) is strong economic growth and a corresponding (low) unemployment rate” (Stanojević and Čehovin Zajc, 2017: 135–136). Several research studies and sources show that the collective ER sys- tems which had sprung up in developed capitalist democracies after WWII began to see rapid changes in the 1980s. Ever since, the power of work- ers’ representative bodies has seen steady declines virtually everywhere – along with the regulatory potential held by collective bargaining and social * Miroslav Stanojević, PhD, Professor, Faculty of Social Sciences, University of Ljubljana, Slovenia; Jožica Čehovin Zajc, PhD, Assistant Professor, Faculty of Social Sciences, University of Ljubljana, Slovenia. Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 661 concertation (Avdagic and Bacaro, 2014). One can thus say that events over the past four decades point to the strong general trend to dismantle col- lective ER systems. Given that the collective ER system is a key feature of democratic capitalist systems that have emerged since WWII, its undoing is a crucial aspect of the massive changes such post-WWII systems, perhaps even the entire post-war social order, have been experiencing. In this article, we compare the collective ER systems in Slovenia, Croatia, Serbia, Hungary and Austria while focusing on the post-crisis period 2013– 2015. Slovenia, Croatia and Serbia once formed part of the specific Yugoslav variant of socialism after WWII; Hungary was more influenced by Soviet cen- tral planning (Whitley, 2000), while Austria built itself after WWII as a state of social partnership in which all features of the collective ER system were developed optimally (Traxler, 1998). During the 1990s, after the collapse of ‘real-socialism’, Hungary sought to resolve the problem of its excessive external debt through a massive sell off of state-owned companies to foreign private owners. The result was the almost complete dominance of foreign capital in the Hungarian economy. The problem of excessive foreign debt in Hungary recurred in the following decade, bringing a radical shift in the electorate towards right-wing politics (Toth et al., 2012). In Slovenia, things developed differently. During the 1990s, Slovenia developed a system simi- lar to that of Austria (and Germany), but it began to change when the coun- try joined the EU and the eurozone. On the other hand, in the 1990s, Croatia and Serbia were involved in the post-Yugoslav wars. In Serbia, which was under international sanctions and then bombed by NATO, the economy was devastated. Croatia’s economic recovery was faster, mainly relying on the development of tourist services. Croatia joined the EU in 2007. Prior to WWI, four of the countries included in our study were part of Austria-Hungary: Austria, Hungary, Slovenia, Croatia, and the north-west- ern part of what is today Serbia. The empire’s centre was in Austria, with Hungarian part being less developed, and Slovenia and Croatia forming its periphery. The differences between the centre of the former empire and its less developed parts continues today and have even increased over time. World Bank data (for 2016) show the following differences in GDP of the countries (in purchasing power terms) in our sample: Austria – USD 50,078; Slovenia – USD 32,885; Hungary – USD 26,681 and Croatia – USD 23,596. Serbia has the lowest GDP per capita of USD 14,512 (World Bank. Database updated 1 July 2017). During the 2013–2015 period, the high unemploy- ment in Serbia and Croatia (levels of between 15% and 20%) constituted the key contextual factor that in principle is hindering the activity of the unions and is incompatible with the optimal operation of collective ER systems. In the same period, the pressure created by unemployment on labour mar- kets in Austria, Slovenia and Hungary was much lower – at a level of below Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 662 10% (Eurostat. Unemployment – national estimate). Hungary has the lowest general union density rate – approximately 10%; in Serbia and Slovenia it is around 20%, and in Austria and Croatia around 30%1 (Industrial Relations in Europe 2014; ETUI (The European Trade Union Institute) – worker partici- pation.eu; Ladjevac, 2017). We can confidently say the big differences seen among the collective ER systems and/or the dynamics of their destabilisation in the countries cov- ered by our study stem from the above-mentioned variations in the level of development achieved and the different historical backgrounds and, even more importantly, from the different pressure created by unemployment and strengths of trade union power. We seek to show that the destabilising mechanisms undermining the col- lective ER systems in those countries are basically similar across the sample. Since unions in certain sectors are similarly proportionally strong regard- less of the national context and, conversely, in other sectors they are mainly weak (or non-existent), and since unionised sectors have been shrinking across the sample and non-unionised ones expanding, one can say the destabilisation of collective ER systems based on sectoral differences, is a common and general outcome of those processes. Analytical framework The Varieties of Capitalism theory (VoC) distinguishes two ideal types of capitalism: a liberal-market economy and a coordinated market economy (Hall and Soskice, 2001). The former type is derived from the American model, and the latter from the (post-war, West) German model. These types are both based on market regulations and the hierarchies incorporated therein, while the latter also makes use of other, non-market regulatory mechanisms, which in essence largely overlap with the collective ER system. Viewed from the VoC theory perspective, the countries studied here are mainly hybrids gravitating towards either of the two types of capital- ism mentioned above. In contrast to Austria, which has the fundamental fea- tures of a coordinated market economy, Slovenia is a post-socialist approxi- mation of that ideal type, Hungary is in between, while Croatia tends to the liberal model (Bohle and Greskovits, 2012). On the other hand, Serbia is a characteristic late-runner, combining typical remnants of the regulation from the socialist period with an ultra-liberal market economy. This rough categorisation in terms of VoC theory should correspond – in accordance with the logic of institutional complementarity (Hall and Soskice, 2001: 1 According to ETUI data, in 2009 the union density in Croatia was 35%. However, because of the high unemployment rate, the union membership among the entire active population is lower. Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 663 17–21) – with the distribution of two sector-specific regimes of employment relations suggested by regulation theory: the system of polyvalent stability based on the specific, hard-to-replace labour force, and the system of market flexibility where the labour force is easy to replace given the non-specific, general skills needed in the sector (Boyer, 2014).2 In the former case, the workers have job stability ensured (standard employment), while in the latter employment is uncertain (non-standard and precarious). In systems which gravitate towards a coordinated market economy, the basic feature of the polyvalent stability systems is job stability – the standard that applies for the lion’s share of active populations. In more liberal market economies, market flexibility is more prevalent. Our analysis uses data collected in CRANET (international survey of HRM policies and practices) research studies,3 which focus on the micro (organi- sational/company) level; these studies’ respondents are HRM managers in companies with 100 or more employees. The subject of research is HRM practices in companies within the three broadly defined sectors: industry, market services, and the public sector. CRANET studies provide two sets of data relevant to our analysis. Based on such data, we are able to establish with great reliability the level of organisations’ numerical flexibility (levels of the liberalisation of dismissal regimes) in the five countries covered by our study (and the three sectors within each one). These data are vital for our analysis since a high level of numerical flexibility (no limits on dismiss- als and employment)4 is incompatible with collective ER systems because it undermines the security of employment and power of the trade unions. Second, the Cranet data enable us to detect variations in the power of workers’ collective representative bodies (trade unions and works’ coun- cils) on the company level. The key indicator is the degree of unionisation on the company level. We start by assuming that in companies with a union density above 50% workers’ representation is strong, meaning the essential ingredient for a collective ER system is present. If less than half the work- ers are union members but there is permanent communication between the union and management, particularly ‘bottom-up communication’, then it is likely the collective ER system includes slight variations (and/or is ‘hybridised’ with HRM practices). If no union is present in the company, or the membership is very low and thus there is no communication with 2 Boyer’s conceptualization is more complex and involves five types of regimes. For the purpose of this analysis, the two above-mentioned types are sufficient. In addition, the nature of data available for this study enables us to analyse only those two, empirically most widely used types. 3 The coordinator of the CRANET network, composed of more than 40 research institutions world- wide, is Cranfield University, School of Management (GB). 4 Those are not accompanied with corresponding safety valves, e.g. high insurance rates in case of unemployment and (professionally and financially strongly supported) active employment policy. Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 664 management, the necessary condition for the presence of a collective ER system is not satisfied. As mentioned in the introduction, our analysis covers the period after the crisis of 2013–2015. Put differently, we compare five countries during the selected (quite specific) segment of a longer time period.5 Below, we first compare the numerical flexibility of companies and then collective rep- resentative bodies in those companies. We then proceed to explain how those two factors interact. In the conclusion, we test our thesis that the sam- ple countries’ collective ER destabilising mechanisms are essentially similar. At this point, let us first briefly recapitulate the results of the analysis of CRANET data for Slovenia covering the decade 2004–2014 (Stanojević and Čehovin Zajc, 2017). It exposed the obvious trend of collective ER systems being rapid dismantled in Slovenian companies. The study showed the pro- cess is very uneven across sectors: it is most powerful in the market ser- vices sector, especially in retail which is based on elemental, non-specific and transferable skills. The market services sector is characterised by the extreme numerical flexibility of companies and correspondingly weak unionisation or their complete absence. The bargaining power of employ- ees in this sector is low (especially in non-unionised parts of the sector). The study also revealed that the undoing of collective ER systems is less intense in the industrial sector in Slovenia. The key groups of employees in those companies possess specific, non-transferable skills. Compared to the market services sector, industrial companies have less numerical flex- ibility. The number of non-unionised companies in this sector is consider- ably lower than in the market services sector. In most companies, union membership is below 50% (at the start of the 2004–2014 decade it exceeded 50%). The analysis showed that unions in industrial companies transformed themselves into representative bodies of groups which possess non-trans- ferable skills. As a result, the situation of those employee groups is quite favourable (see Stanojević and Čehovin Zajc, 2017). Numerical flexibility As mentioned in the introduction, the sample countries may be divided into two distinct groups regarding unemployment: in Serbia and Croatia, the unemployment rate ranges from 15% to 20%, while in Austria, Hungary and Slovenia it is much lower – below 10 percent. 5 In other words, our study is not a comparison of the trajectories of collective ER regime changes in the examined countries. Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 665 Figure 1: UNEMPLOYMENT RATE (% OF TOTAL LABOUR FORCE, NATIONAL ESTIMATE) Source: The World Bank (2018): World Development Indicators. Unemployment – national estimate. Within this general framework marked by characteristic differences in the pressure created by ‘the reserve army of the unemployed’, figures show- ing dismissals and new employment dynamics on the company level are presented below. As Table 1 shows, in Austria and Hungary during the 2013–2015 period the number of employees rose in almost half the companies (47.8% in Austria, 45.9% in Hungary), while it declined in approximately one in four organisations (25.4% and 23.0%). In Croatia and Slovenia, employee numbers fell in almost half of all com- panies (45.6% and 45.2%), while in slightly more than one-third of them it increased (35.7% and 36.7%). These trends make Croatia and Slovenia a mir- ror image of Austria and Hungary. Serbia is a case in its own right. In half of all Serbian companies, the num- ber of employees remained the same (51.8%); approximately one-quarter of companies (23.7%) downsized their workforces, and another one-quarter (24.6%) increased them. Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 666 Table 1: CHANGE IN EMPLOYEE NUMBERS IN PAST THREE YEARS Change in no. of employees Country 1 – decreased 2 – not changed 3 – increased Austria 25.4% 26.8% 47.8% Hungary 23.0% 31.1% 45.9% Serbia 23.7% 51.8% 24.6% Croatia 45.6% 18.7% 35.7% Slovenia 45.2% 18.1% 36.7% Sector 1 – decreased 2 – not changed 3 – increased Austria industry 18.3% 28.0% 53.7% market services 32.3% 24.6% 43.1% public services 27.4% 27.4% 45.2% Hungary industry 19.3% 27.5% 53.2% market services 18.9% 31.6% 49.5% public services 37.3% 32.2% 30.5% Serbia industry 30.0% 46.0% 24.0% market services 18.4% 42.1% 39.5% public services 20.0% 76.0% 4.0% Croatia industry 50.8% 23.1% 26.2% Market services 38.2% 13.2% 48.5% public services 50.0% 21.1% 28.9% Slovenia industry 40.0% 16.9% 43.1% market services 52.8% 16.7% 30.6% public services 45.5% 25.5% 29.1% Local / global market 1 – decreased 2 – not changed 3 – increased Austria local 27.6% 26.8% 45.5% global 22.7% 26.8% 50.5% Hungary local 23.9% 36.5% 39.6% global 20.0% 18.9% 61.1% Serbia local 30.8% 53.8% 15.4% global 8.6% 45.7% 45.7% Croatia local 48.5% 15.8% 35.6% global 41.5% 23.1% 35.4% Slovenia local 48.9% 21.7% 29.3% global 38.1% 17.5% 44.4% Source: Cranet, 2015. In Austria, in all three sectors – industrial, market services, and public sec- tor – most companies increased their employee numbers (around half of all companies in each sector, with industrial sector companies leading the way with a 53.7% share). In Hungary, the proportions of companies in the indus- trial and market services sectors that increased their employee numbers Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 667 are similar – 53.2% of companies in the industrial sector and 49.5% in the market services sector. In the public sector, the share of companies that downsized their workforce exceeds the share of companies that increased it (37.3% compared to 30.5%). In Croatia, half the organisations in the industrial (50.8%) and public sec- tors (50.0%) reduced their employee numbers (in one-fifth of organisations the numbers did not change). Half the Croatian companies in the market services sector (48.5%) increased their employee numbers. In Slovenia, the proportion of companies which expanded employment is not significantly high in any of the three sectors. In the industrial sector, the shares of com- panies increasing and cutting their workforces are in balance (40.0% and 43.2%, respectively), while the market services and public sectors have the biggest shares of companies that reduced their employee numbers (approx- imately half of them, or 52.8% in the market services sector and 45.5% in the public sector). Serbia has the highest percentage of companies making no change in employee numbers. This is obvious in the public sector where most organi- sations – three-quarters of them (76.0%) – did not change the number of their employees. The market services sector has the greatest share of com- panies that increased their employee numbers (two-fifths of them; 39.5%). In the industrial sector, 46.0% of companies kept the number of employ- ees the same, companies downsizing their workforce accounted for one- third (30.0%) of all companies, while those that expanded it for one-quarter (24.0%) of companies. To recapitulate, the sectoral comparison showed that in Austria most companies increased their workforce in each of the three sectors, with industrial companies leading the way. In Hungary, the situation is similar in the industrial and market services sector, while the drop in employee num- bers is more conspicuous in the public sector. In Slovenia, the situation is more balanced in all three sectors – the proportions of industrial companies increasing and of those decreasing their workforce are roughly equal, while in the market services and the public sector slightly more companies cut their workforce. In Croatia, the workforce rose in companies in the market services sector, while in the industrial and public sectors the proportion of companies shrinking their employee numbers is larger. Serbia has a rela- tively big share of companies in the market services sector that increased their workforce, although the industrial and public sectors saw employment stagnation. These comparisons show that employment in Austria was increasing in all sectors, especially in the industrial sector. In Croatia and Serbia, employ- ment in the industrial sector and the public sector was stagnant or in some cases decreased, but employment in the market services sector showed a Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 668 robust increase. Slovenia and Hungary are ‘in between’ these two poles, each in their own way. In Austria and Hungary, during the period 2013–2015, the biggest employment increase was seen for organisations oriented to the global mar- ket (the export sector) – in Austria, half the companies (50.5%) that increased employment come from the export sector, and in Hungary this proportion is even higher at slightly below two-thirds of companies (61.1%). Collective representation of workers We have already mentioned that in the countries under scrutiny the union density rates vary, ranging from some 10% in Hungary and 20% in Serbia and Slovenia to some 30% in Austria and Croatia. Figure 2: TRADE UNION DENSITY RATES Source: OECD (Organisation for Economic Co-operation and Development) (2018); Trade union density in OECD countries. Union density on the company level Comparing the union densities on the company level reveals two extremes: Austria has the smallest share of companies in which unions are not present (10.7%), whereas the share of zero-unionisation companies in Hungary is 50.2%. In both countries, the union membership is below 50% in Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 669 roughly half of all companies. In Austria, union membership exceeds 50% in two-fifths of companies (37.7%), while in Hungary companies with such high union membership virtually do not exist (only 3.2%). Compared to the extremes mentioned above, the companies in Croatia and Slovenia are closer to Austria than Hungary. Although both countries have higher proportions of companies with no unions present than Austria (16.9% of companies in Croatia and 14.2% in Slovenia), those proportions are much lower than in Hungary. In both countries, companies with union membership below 50% constitute the largest proportions (46.6% of com- panies in Croatia and 60.1% of companies in Slovenia). In one-third (36.5%) of Croatian companies and one-quarter (25.7%) of Slovenian companies, union densities exceed 50%. In Croatia, the proportions of companies with less than 50% union density and those with over 50% are similar to those in Austria. Slovenia has the biggest share of companies – 60.1% – with low union membership (below 50%). Serbia stands between these two extremes (Austria and Hungary). In one-third (30.8%) of Serbian companies, unions are not present; in one-fifth (19.5%) of companies, the union density is below 50%, and in one-half of companies (49.7%) it exceeds 50%. Austria has the highest union membership in the public sector and the industrial sector – half (50.9%) of public-sector companies and two-fifths (41.3%) of industrial companies have union membership exceeding 50%. In the market services sector, union membership is low (less than 50%) in most companies (71.1%). Further, the same sector in Austria is characterised by the largest share of companies (15.8%) in which trade unions cannot be found at all. In Hungary, unions are non-existent in (almost) two-thirds of companies in the industrial sector (57.3%) and the market services sector (56.3%), and in one-third (29.8%) of public-sector companies. Unions are present in two- thirds (68.1%) of public-sector companies, but the union density is low in all cases (less than 50%). Unions are not present in one-third of companies in the market services sector in Slovenia and Croatia (30.6% and 31.6%). In Croatia, the industrial sector is dominated by highly unionised companies (slightly less than two- thirds or 57.6% of all companies). On the other hand, the Slovenian indus- trial sector is characterised by a similar share of companies (60.7%) with low union membership. In Croatia, the share of public companies with high union membership is greater than in Slovenia (34.4% vs. 26.0%). In the industrial and public sectors in Slovenia, the proportions of companies with low union density are typically higher than in Croatia. In Serbia, slightly less than two-thirds (57.4%) of companies in the market services sector and one-fifth of companies (21.9%) in the industrial sector Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 670 have no union presence. On the other hand, in the industrial and public sec- tors, companies with high union membership dominate (around two-thirds of them, or 58.9% of companies in the industrial sector and 59.4% of compa- nies in the public sector). Table 2: UNION MEMBERSHIP Proportion of Country 0% 1–50% 51–100% Austria 10.7% 51.6% 37.7% Hungary 50.2% 46.5% 3.2% Serbia 30.8% 19.5% 49.7% Croatia 16.9% 46.6% 36.5% Slovenia 14.2% 60.1% 25.7% Sector Austria industry 7.9% 50.8% 41.3% market sector 15.8% 71.1% 13.2% public services 10.5% 38.6% 50.9% Hungary industry 57.3% 38.5% 4.2% market sector 56.3% 40.8% 2.8% public services 29.8% 68.1% 2.1% Serbia industry 21.9% 19.2% 58.9% market sector 57.4% 11.1% 31.5% public services 6.3% 34.4% 59.4% Croatia industry 6.8% 35.6% 57.6% market sector 31.6% 52.6% 15.8% public services 9.4% 56.3% 34.4% Slovenia industry 8.2% 60.7% 31.1% market sector 30.6% 52.8% 16.7% public services 10.0% 64.0% 26.0% Local / global market Austria local 5.6% 50.0% 44.4% global 17.6% 52.9% 29.4% Hungary local 46.0% 50.0% 4.0% global 63.0% 34.6% 2.5% Serbia local 26.4% 18.2% 55.5% global 40.8% 22.4% 36.7% Croatia local 18.2% 50.0% 31.8% global 15.5% 41.4% 43.1% Slovenia local 15.1% 59.3% 25.6% global 13.3% 60.0% 26.7% Source: Cranet, 2015. Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 671 To sum up, the comparison shows the three sectors in the countries stud- ied here have certain key similarities. As a rule, trade unions are weakest in the market services sector; all countries have the biggest shares of zero- unionisation companies in this sector. Unions are much stronger in the industrial and public sectors than in the market services sector in all coun- tries in the sample, except Hungary where the share of industrial companies with weak unions (or union absence) is as low as in the market services sector. In two-thirds (63.0%) of export companies in Hungary, and two-fifths (40.8%) of export companies in Serbia, trade unions are not present. On the other hand, in Hungary, one-third (34.6%) of export companies have low union density, while virtually no company encounters high union density. In Serbia, slightly more than one-third of companies in the export sector (36.7%) have high union density. In Croatia too, union density in the export sector is quite high (43.2%). In Austria and Slovenia, the export sector is dominated by companies with low union membership (52.9% and 60.0% compared to 29.4% and 26.7% of companies with high union membership). The same sector in Austria also has a relatively large share of companies (17.6%) without any unions present. Works’ councils Table 3: DO YOU HAVE A JOINT CONSULTATIVE COMMITTEE OR WORKS’ COUNCIL? (% WITHIN SECTOR, POSITIVE ANSWERS) % within sector (YES) Austria Hungary Serbia Croatia Slovenia Yes – Total 82.7% 35.2% 28.8% 60.0% 58.2% Sector industry 79.0% 35.5% 36.1% 76.9% 69.2% market services 75.8% 32.3% 17.3% 49.3% 55.6% public services 93.1% 43.1% 31.3% 50.0% 45.5% Market local 90.2% 35.2% 27.8% 55.4% 54.9% global 73.2% 34.4% 31.3% 67.7% 62.5% Source: Cranet, 2015. In Austria where – like in Germany – the works’ councils are the only collective representative body for workers within organisations,6 works’ councils are present in four-fifths (82.7%) of companies. In Hungary and the three post-Yugoslavia states, the collective representation of workers 6 In this system, union trustees and members are present within individual organizations, but they are not organized into autonomous units – organizational trade union. Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 672 – where it exists – is usually dual: in addition to trade unions, works’ coun- cils are also present in most companies. In Hungary and Serbia, works’ councils exist in one-third (35.2% and 28.8%) of all companies, and in Croatia and Slovenia in two-thirds (60.0% and 58.2%). With the exception of Serbia, in all the other countries, works’ councils are more characteristic of larger organisations that have more than 250 employees. In the three post-Yugoslavia states, they are mainly found in industrial organisations – in three-quarters (76.9%) of all industrial com- panies in Croatia, in two-thirds (69.2%) of them in Slovenia. Consistent with those differences, in the former Yugoslav states works’ councils are more likely to be found in export-oriented companies than in local-market-ori- ented companies. Top-down and bottom-up communication via workers’ representative bodies Interaction or communication between the workers and the manage- ment (top-down and bottom-up communication) through workers’ repre- sentative bodies is important information revealing the importance of such representative bodies. We established that such communication is present in virtually all highly unionised companies (with higher than 50% union densities), while in companies with very low union densities (10% or lower) interactions between workers’ representative bodies and managements are more the exception than the rule. Top-down communication via workers’ representative bodies Top-down communication through trade unions representatives: Top- down communication between employees and management through unions is weak in Austria and Hungary. In fact, in most companies (74.9% in Austria and 71.9% in Hungary) such communication does not exist. However, there is a significant systemic difference explaining this apparent similarity. In the Austrian system of industrial relations, the key representa- tive body of employees in a company is the works’ council. As a result of this feature, trade unions do not act as an interaction conduit linking workers and management. In Hungary, on the other hand, workers’ representative bodies – both trade unions and works’ councils – are weak or non-existent. Accordingly, their role in employee–management interactions is marginal. In the three post-Yugoslavia states, unions are an important communica- tion channel used in three-fifths (60.1%) of companies in Serbia and two- thirds of companies in Croatia and Slovenia (68.7 % and 72.3%). Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 673 d) WORKS COUNCIL % within sector (NO) Austria Hungary Serbia Croatia Slovenia Total Not at all – 20.8% 70.3% 81.8% 43.3% 43.1% Sector industry 23.8% 64.5% 74.0% 24.6% 32.3% marketing 28.1% 75.8% 88.9% 53.3% 38.2% public services 11.3% 73.2% 87.5% 58.3% 60.4% Company size up to 250 employees 38.2% 84.6% 83.5% 61.3% 52.9% over 250 employees 16.8% 54.3% 79.0% 39.2% 30.9% Share of graduates up to 25% 20.7% 61.4% 78.4% 41.8% 31.5% 26% and above 15.4% 74.1% 85.7% 49.1% 60.7% Market local market 14.0% 71.3% 83.6% 47.3% 45.5% global market 29.8% 70.2% 77.6% 36.7% 39.7% Without TOP-DOWN communication with employees through a) TRADE UNION representatives % within sector (NO) Austria Hungary Serbia Croatia Slovenia Not at all – Total 74.9% 71.9% 39.9% 31.3% 27.7% Sector industry 79.0% 69.1% 35.6% 14.5% 16.9% market services 83.1% 77.9% 58.5% 47.7% 38.2% public services 62.5% 67.8% 18.8% 30.6% 35.2% Company size up to 250 employees 82.9% 83.3% 49.0% 41.2% 39.1% over 250 employees 72.9% 58.9% 25.8% 28.1% 13.2% Share of graduates up to 25% 73.6% 68.2% 42.0% 29.0% 20.9% 26% or over 77.5% 73.3% 37.7% 37.9% 40.0% Market local market 66.4% 72.0% 34.9% 33.3% 32.2% global market 85.4% 76.8% 51.0% 28.6% 22.6% b) through WORKS COUNCIL % within sector (NO) Austria Hungary Serbia Croatia Slovenia Total Not at all – 26.0% 70.8% 80.3% 43.4% 44.7% Sector industry 25.9% 67.3% 72.6% 24.6% 33.8% marketing 35.4% 70.5% 90.4% 54.1% 45.5% public services 18.1% 78.6% 81.3% 56.8% 59.6% Company size up to 250 employees 31.4% 84.1% 81.1% 57.6% 52.9% over 250 employees 24.3% 56.3% 79.0% 40.0% 34.3% Share of graduates up to 25% 25.2% 67.1% 79.5% 40.8% 31.5% 26% or over 25.0% 72.2% 80.9% 50.0% 66.1% Market local market 21.3% 71.7% 80.6% 48.9% 46.0% global market 32.3% 70.5% 79.6% 34.4% 43.5% Without BOTTOM-UP communication through c) TRADE UNION representatives % within sector (NO) Austria Hungary Serbia Croatia Slovenia Total Not at all – 70.8% 71.4% 39.6% 30.4% 21.4% Sector industry 76.3% 70.6% 32.9% 17.7% 10.8% marketing 80.3% 75.8% 57.4% 42.9% 29.4% public services 55.6% 66.7% 25.0% 30.6% 29.6% Company size up to 250 employees 80.6% 86.8% 46.4% 39.4% 33.7% over 250 employees 68.3% 54.3% 29.0% 27.6% 5.9% Share of graduates up to 25% 72.7% 65.5% 38.6% 29.0% 15.4% 26% or over 62.5% 75.1% 41.4% 35.7% 31.7% Market local market 63.1% 70.0% 36.4% 29.5% 25.6% global market 80.2% 78.7% 46.9% 32.3% 16.1% Table 4: NO COMMUNICATION WITH EMPLOYEES VIA WORKERS' REPRESENTATIVE BODIES Source: Cranet, 2015. Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 674 Most market service companies in Austria (83.1%) and Hungary (77.9%) do not use trade unions as a communication channel. In Austria, the reason is the men- tioned feature of the system while, in Hungary, it is the absence of unions or their marginal role. In Serbia, slightly more than half the companies (58.5%) do not com- municate with workers through unions, and in Croatia the situation is the same in slightly less than half of all companies (47.7%). In the market sector in Slovenia, two-fifths of all companies (38.2%) do not communicate through unions. Trade unions are mainly used as a top-down communication conduit in larger companies (with more than 250 employees), while this practice is less present in smaller companies. In Austria and Serbia, it is more used in public-sector companies, and in Croatia and Slovenia it is the most common way of top-down communication in all sectors (used in approximately 70% of all companies). In Croatia and Slovenia, trade unions are used as a top- down communication channel mainly in export-oriented companies, and in Austria and Serbia in companies focused on the domestic market. Top-down communication through works’ council: We again note that in Austria the key institution for workers’ representation in companies is the works’ council. In accordance with such arrangement, managements mainly communicate with the workers in this way (in around three-quarters or 74.0% of companies). In Hungary and Serbia, this communication channel is the least used (in around one-third or 29.2% of companies in Hungary and one-fifth or 19.7% of companies in Serbia). In Croatia and Slovenia, works’ councils serve as the top-down communication channel in more than half the companies (56.6% in Croatia and 55.3% in Slovenia). In Austria, this top-down communication model is most used in the public sector, and in Croatia and Slovenia in the industrial sector. In the market services sector in Serbia, works’ councils are a relatively insignificant communication channel (90.4% of companies do not use it), with a similar situation being observed in Hungary where the figure is 70.5%. In the market services sector in Croatia and Slovenia, it is used by about half the companies (45.9% in Croatia and 54.5% in Slovenia). A coun- try comparison shows that in the market services sector this communication channel is prevalent in Austria (used by 64.6% of companies), but this share is lower than in other sectors in Austria. In Austria, works’ councils as a top-down communication channel are more used in companies that focus on the domestic market (78.7% of them), while in Croatia they are chiefly used in export-oriented companies (65.6%). Bottom-up communication via workers’ representative bodies Bottom-up communication through trade union representatives: Due to the mentioned inherent feature of the Austrian system and the weak unions Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 675 in Hungary, in those two countries the bottom-up communication through unions almost does not exist in the lion’s share of companies (70.8% in Austria and 71.4% in Hungary). By contrast, in all three former Yugoslav states, this mode of communication is found in most companies (60.4% in Serbia, 69.6% in Croatia and 78.6% in Slovenia). The bottom-up communication via trade unions in Austria, Hungary and in Serbia is more practised by public companies than industrial companies. In Slovenian industrial companies, bottom-up communication through unions occurs in nearly 90% of companies, and in Croatia in more than 80%. Bottom-up communication via trade unions is least used in the market ser- vices sector: it exists in only one-quarter (24.2%) of market services compa- nies in Hungary, in two-fifths (42.6%) of such companies in Serbia, more than half of such companies (57.1%) in Croatia, and two-thirds (70.6%) in Slovenia. This mode of communication is more characteristic of the larger compa- nies in our sample, and is most widely used in Slovenia where just 5.9% of companies with more than 250 employees do not use it. With the exception of Slovenia, in the other countries in the sample this mode of communication is less used by export companies than by companies focused on the domestic market. In Slovenia, it is used by 83.9% of all companies in the export sector. Bottom-up communication through works council: In Austria, works’ councils constitute an important bottom-up communication channel – one used by four-fifths (79.2%) of all companies. This communication channel is also found in the majority of companies in Slovenia (56.9%) and Croatia (56.7%); in Hungary, it is used by (only) one-third of companies (29.7%) and in Serbia by one-fifth (18.2%). Bottom-up communication through works’ councils occurs in nearly 90% of public-sector companies in Austria, but only in 10% of public-sector companies in Serbia. In the industrial sector, this communication channel is most widely used in Austria (by around three-quarters of companies, or 79.2% of them) and in Croatia (75.4%). It is least used in Serbia where 74.0% of all companies do not use it, and in Hungary where 64.5% of companies do not use it. In the Slovenian industrial sector, bottom-up communication via works’ councils occurs in around two-thirds (67.7%) of companies. Bottom-up communication through works’ councils is not present in the large majority of market service companies in Serbia (88.9%) and Hungary (75.8%). In Croatia, this channel of communication is used in roughly half of companies in the market services sector (46.7%), and in Slovenia and Austria by the majority of market service companies (61.8% and 71.9%), although in Austria the share for the market services sector is smaller than for any other sector in this country. In all countries, bottom-up communication through works’ coun- cils occurs more often in larger companies; in Austria it is used more in Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 676 companies that focus on the domestic market, and in Croatia and Slovenia in export-oriented companies. Trade unions and employment – sector differences In Austria, the most developed country in our sample with the lowest unemployment rate, the number of workers increases more in de-union- ised than in unionised companies. In unionised and de-unionised compa- nies from the industry, the figures on workforce expansion/reduction are close to each other; in the market service sector, the workforce fluctuation is greater in de-unionised companies; and in the public sector half of non- unionised organisations (50%) increased and half of unionised companies (45.20%) reduced their employee numbers. Accordingly, the process of dis- mantling the collective ER regimes in Austria is indeed underway but, when viewed comparatively, it is gradual or non-intense. In Slovenia, the share of organisations cutting their employee numbers – compared to the share of companies expanding their workforce, was higher in both unionised and de-unionised companies, but new employment was less intensive among unionised companies. In the market services sector, almost two-thirds of unionised companies reduced their employee num- bers. In industry, new employment was seen in two-thirds (62.50%) of non- unionised companies, but – according to the small number of non-union- ised companies in industry – the tendency is weak. It can be said that the basic trend of redeployment, which is undermining the trade unions and (thus) bringing the erosion of collective ER systems, is obvious in Slovenia as well. The trend is definitely stronger in Slovenia than in Austria, although the overall ‘architecture’ of the collective ER systems in Slovenia is not (yet) threatened by such redeployment. In Croatia and Serbia, both of which have been feeling the general pressure of unemployment, the number of employees has been rapidly increasing in the market services sector where trade unions are the weak- est. Therefore, it could be said that developments in the market services sec- tor triggered the process of the dismantling of collective ER. The process is already underway in Croatia and is imminent in Serbia. Croatia is halfway through radically dismantling its collective ER systems. Serbia, which for the time being maintains the relations it inherited from the previous system in the public sector and parts of the industrial sector and is experiencing a rapid rise in employment in the market services sector, is at the beginning of the dismantling process. In Hungary, where trade unions are weak in all sectors (and absent in 63% of companies in the export sector dominated by foreign capital – see Table 2), the process of dismantling the collective ER system is virtually Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 677 Table 5: DIFFERENCES AMONG SECTORS IN WORKFORCE EXPANSION/ REDUCTION BY UNIONISED AND DE-UNIONISED COMPANIES de-unionised unionised country sector de-creased not changed increased de- creased not changed increased Austria Total 22.10% 27.30% 50.60% 31.70% 27.00% 41.30% industry 18.30% 26.70% 55.00% 15.00% 35.00% 50.00% market services 34.00% 20.80% 45.30% 25.00% 41.70% 33.30% public services 12.50% 37.50% 50.00% 45.20% 16.10% 38.70% Hungary Total 20.10% 26.50% 53.40% 29.90% 42.90% 27.30% industry 15.80% 25.00% 59.20% 25.80% 35.50% 38.70% market services 13.90% 26.40% 59.70% 34.80% 47.80% 17.40% public services 39.50% 26.30% 34.20% 33.30% 42.90% 23.80% Serbia Total 17.80% 42.20% 40.00% 29.20% 58.50% 12.30% industry 35.30% 35.30% 29.40% 28.10% 53.10% 18.80% market services 8.30% 41.70% 50.00% 38.50% 46.20% 15.40% public services 75.00% 25.00% 25.00% 75.00% Croatia Total 21.60% 17.60% 60.80% 55.60% 19.70% 24.80% industry 27.30% 72.70% 61.10% 22.20% 16.70% market services 27.60% 13.80% 58.60% 44.40% 13.90% 41.70% public services 27.30% 18.20% 54.50% 59.30% 22.20% 18.50% Slovenia Total 45.70% 14.30% 40.00% 45.00% 21.70% 33.30% industry 37.50% 62.50% 40.40% 19.30% 40.40% market services 45.50% 9.10% 45.50% 58.30% 20.80% 20.80% public services 50.00% 25.00% 25.00% 42.10% 26.30% 31.60% Total Total 22.60% 26.40% 51.10% 41.00% 31.00% 28.10% industry 18.60% 25.60% 55.80% 39.20% 29.90% 30.90% market services 22.80% 23.80% 53.40% 42.60% 29.60% 27.80% public services 28.40% 31.20% 40.40% 42.30% 32.80% 24.80% * The de-unionised group comprises organisations without trade union presence, meaning that none of their employees is a union member, as well as organisations where trade uni- ons are formally present but are inactive and do not play any role in intra-company commu- nications. Source: Cranet, 2015. Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 678 complete since a collective ER system cannot survive without trade unions being present. Accordingly, the current absence of strong unemployment pressure in Hungary is inconsequential for the dismantling of the country’s collective ER system. Conclusion In the introduction to this article, we emphasised that the mechanisms behind the destabilisation of collective ER systems are in essence similar across the sample. We established that the pressure of unemployment (viewed compara- tively) is especially strong in two countries in the sample, namely Serbia and Croatia, that it is significantly lower in Slovenia, and less consequential and/ or inconsequential in Hungary and Austria. The analysis also showed that in all countries in our sample trade unions are stronger in the industrial and public sectors, and weakest in the market services sector. Accordingly, this sector has the biggest proportion of de-unionised organisations (those in which trade unions are not present and/or have no influence whatsoever). The comparison of workforce fluctuations in unionised and de-union- ised companies in five countries clearly shows the general trend of a rise in the share of de-unionised organisations which are increasing their work- force on one hand, and the decline and stagnation of employment levels in unionised organisations on the other. The former group is concentrated in the market services sector, and the latter in the industrial and public sectors. This kind of redeployment, which firmly encourages de-unionisation and the dismantling of collective ER systems in general became a (more or less) prevalent trend during the post-crisis period 2013–2015 in the countries covered by this study. The differences in the intensity of dismantling collective ER systems, from incremental (Austria and Slovenia) to radical (Croatia and Serbia) and completed processes (Hungary) raise a new set of research questions. Which type of regulation is to replace or has already replaced the collec- tive ER regimes? A general answer is that the liberal market economy under- mines regulatory mechanisms of the coordinated market economy, that col- lective employment relations are being replaced by individual employment relations and that, accordingly, institutions and processes of workers’ co- management are being reduced to human resources management. Authors who researched HRM regimes established notions of qualita- tively different management models, namely hard and soft HRMs (Sisson and Storey, 2000; Herry and Noon, 2001). Our analysis shows the dynamics of the processes that are eroding the collective ER differ among both countries and sectors. Accordingly and noting the dual typology of HRM regimes, one of Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 679 the findings of our analysis is that in countries and sectors with weak collec- tive ER regimes, or without any collective ER regimes, the practices of hard HRM are more likely to replace collective ER regimes and, conversely, mod- erate collective ER dismantling processes are being matched with hybridisa- tion practices ranging from moderate, soft HRM practices to more or less modified collective ER regimes (combined with elements of HRM practices). BIBLIOGRAPHY Avdagic, Sabina and Lucio Bacaro (2014): The Future of Employment Relations in Advanced Capitalism: Inexorable Decline? In Adrian Wilkinson, Geoffrey Wood and Richard Deeg (eds.), The Oxford Handbook of Employment Relations, Comparative Employment Systems, 701–725. Oxford: Oxford University Press. Bohle, Dorothee and Bela Greskovits (2012): Capitalist Diversity on Europe’s Periphery. Ithaca, NY: Cornell University Press. Boyer, Robert (2014): Development and Extension of ‘Regulation Theory’ and Employment Relations. In Adrian Wilkinson, Geoffrey Wood and Richard Deeg (eds.), The Oxford Handbook of Employment Relations, Comparative Employment Systems, 114–155. Oxford: Oxford University Press. Hall, Peter A. and David Soskice (2001): An Introduction to Varieties of Capitalism. In Peter A. Hall and David Soskice (eds.), Varieties of Capitalism, The institu- tional foundations of Comparative Advantage, 1–68. Oxford: Oxford University Press. Herry, Edmund and Mike Noon (2001): A Dictionary of Human Resource Management. Oxford: Oxford University Press. Ladjevac, Bojan (2017): Trade Unions in Srbija on the Move? Berlin: Friedrich-Ebert- Stiftung. Sisson, Keith and John Storey (2000): The Realities of Human Resources Management, Managing the Employment Relationship. Buckingham: Open University Press. Stanojević, Miroslav and Jožica Čehovin Zajc (2017): Kadrovske politike v pogo- jih demontaže kolektivnih delovnih razmerij (HRM policies in the circum- stances of dismantling the collective employment relations). In Andrej Kohont and Miroslav Stanojević (eds.), Razpotja in prelomi: spremembe na področju menedžmenta človeških virov v Sloveniji (Crossroads and Shifts: the changes in the HRM field in Slovenia), 133–162. Ljubljana: Založba FDV. Tóth András, Neumann László and Hosszú Hortenzia (2012): Madžarska’s Full- blown Malaise. In Steffen Lehndorf (ed.) A Triumph of Failed Ideas: European Models of Capitalism in the Crisis, 137–153. Brussels: ETUI. Traxler, Franz (1998): Austria: Still the Country of Corporatism. In Anthony Ferner and Richard Hyman, Changing Industrial Relations in Europe (second edition), 239–261. Oxford: Blackwell. Whitley Richard (2000): Divergent Capitalisms: The Social Structuring and Change of Business Systems. Oxford: Oxford University Press. Miroslav STANOJEVIĆ, Jožica ČEHOVIN ZAJC TEORIJA IN PRAKSA let. 56, 2/2019 680 SOURCES Cranet (2015): Cranet – international survey of HRM policies and practices. Cranfield: University of Cranfield (database). ETUI – The European Trade Union Institute (2018): The gateway to information on worker participation issues in Europe. Accessible at https://www.worker- participation.eu/, 14. 6. 2018. European Commission (2015): Industrial Relations in Europe 2014. Luxembourg: Publications Office of the European Union. OECD (2018): Trade union density in OECD countries. Accessible at: https://stats. oecd.org/Index.aspx?DatasetCode=STLABOUR, 14. 6. 2018. The World Bank (2018): World Development Indicators. Unemployment – national estimate. Accessible at https://data.worldbank.org/indicator/SL.UEM.TOTL. NE.ZS?view=chart, 21. 5. 2018. The World Bank (2018): World Development Indicators. Accessible at: https://data. worldbank.org/, 14. 6. 2018