German Works Councils Old and New: Incidence, Coverage and Determinants

Addison, John T.; Bellmann, Lutz; Schnabel, Claus; Wagner, Joachim Article German Works Councils Old and New: Incidence, Coverage and Determinants Schmollers Jahrbuch – Zeitschrift für Wirtschaftsund Sozialwissenschaften. Journal of Applied Social Science Studies Provided in Cooperation with: Duncker & Humblot, Berlin Suggested Citation: Addison, John T.; Bellmann, Lutz; Schnabel, Claus; Wagner, Joachim (2003) : German Works Councils Old and New: Incidence, Coverage and Determinants, Schmollers Jahrbuch – Zeitschrift für Wirtschaftsund Sozialwissenschaften. Journal of Applied Social Science Studies, ISSN 1865-5742, Duncker & Humblot, Berlin, Vol. 123, Iss. 3, pp. 339-358, https://doi.org/10.3790/schm.123.3.339 This Version is available at: https://hdl.handle.net/10419/292060 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Schmollers Jahrbuch 123 (2003), 339-358 Duncker & Humblot, Berlin German Works Councils Old and New: Incidence, Coverage and Determinants* By John T. Addison, Lutz Bellmann, Claus Schnabel, and Joachim Wagner Abstract Although works councils are a core element of the German system of industrial rela tions, there is little reliable information on their incidence and coverage. This paper uses data from the nationally representative IAB establishment panel to fill this gap. We examine the frequency of works councils by establishment size and broad sector for eastern and western Germany, while at the same time charting the determinants of their presence. Furthermore, we identify newly established works councils and the circum stances of their formation. Finally, we discuss the consequences of our findings for eco nomic analysis of the institution and for public policy. Zusammenfassung Obgleich die Betriebsräte ein Kernelement des bundesdeutschen Systems der in dustriellen Beziehungen darstellen, gibt es kaum zuverlässige Informationen über ihre Verbreitung. Der folgende Beitrag verwendet Daten des IAB-Betriebspanels um die Verbreitung von Betriebsräten in unterschiedlichen Wirtschaftszweigen und Betriebs größenklassen in Ostund Westdeutschland zu dokumentieren sowie die Bestimmungs gründe für die Existenz einer Arbeitnehmervertretung zu identifizieren. Darüber hinaus fragen wir nach den Gründen für die erstmalige Wahl eines Betriebsrates in bis dahin vertretungslosen Unternehmen und diskutieren die Konsequenzen unserer Befunde für die zukünftig anstehende ökonomische Analyse rechtlich autorisierter Arbeitnehmer vertretungen. JEL Classification: 150 * This study uses data from the IAB establishment panel, which is financially sup ported by the European Social Fund. We are indebted to Holger Alda of the IAB for his help with the data. An earlier version of this paper was presented at a workshop of the University of Witten/Herdecke in February 2002. We thank the participants for their helpful remarks. We also benefited from the comments of two anonymous referees of this journal. In each case, the usual disclaimers apply. Schmollers Jahrbuch 123 (2003) 3 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.3.339 | Generated on 2023-04-04 12:32:11 340 John T. Addison, Lutz Bellmann, Claus Schnabel, and Joachim Wagner 1. Motivation After protracted and highly controversial debate, the German Bundestag passed the law reforrning the pre-existing Works Constitution Act (Betriebs ve,fassungsgesetz) on June 22, 2001. The new Works Constitution Reform Act (BetrVe,f-Reformgesetz) was approved by the Bundesrat on July 13, 2001, and becarne effective on July 28, 2001, the day after its announcement in the Federal Law Bulletin (Bundesgesetzblatt).1 Section 1 of both the old and the new law provide for the election of works councils in all establishments that normally have five or more employees with voting rights, including three who are eligible to be works councilors. (Voting rights accrue to all workers aged 18 years and above, and eligible employees are all workers who have been employed by the firm for at least six months.) Although leaving this particular size threshold unchanged, the new legislation seeks to facilitate works council formation through simplified election proce dures, inter al. This goal reflects the lirnited penetration of works councils. Among small establishments, works council plants are a distinct minority de spite the low employment size criterion and strict sanctions against employers who attempt to interfere with the process of works council formation. Note also that the new legislation does make a large number of formal changes that have the effect of increasing the power and authority of works councils, with the stated aim of improving the econornic performance of the institution. In short, the central aim of the German govemment in framing the legisla tion was to increase the number (and reach) of works councils and thereby shrink the size of what was termed a 'codeterrnination-free zone' by the Kom mission Mitbestimmung (Codetermination Comrnission), a high-level group of experts set up to evaluate, among other things, the experience with the work ings of the old Works Constitution Act of 1972.2 According to the Commis sion's final report, issued in 1998, no less than 60 percent of all private-sector employees worked in firms without a works council (and without employee representatives on the supervisory board). Moreover, the Comrnission pre dicted a rise in this share for structural reasons; in particular, the growing im1 There is as yet no official English-language version of the law. The Works Consti tution Reform Act, documenting changes between the old and the new law, is published in Deutscher Bundestag, 14. Wahlperiode, Drucksache 14 / 57 41 vom 2. April 2001, and is available on the web at (www.bundestag.de). The new Act is published in stand alone form in the Bundesgesetzblatt BGBl. 1 S. 1852, and is available on the web at (www.bma.bund.de/ download/ gesetze/BetrVG.pdf). For a review of the evolution of workplace codetermination in Germany, the innovations of the new law, and the contro versies over costs and benefits of works councils, see Addison, Bellmann, Schnabel, and Wagner (2003). 2 See Kommission Mitbestimmung (1998). An English-language summary of the final report of the Commission is available at (www.mpi-fg-koeln.mpg.de/ endbericht/ inhalt_e.html). Schmollers Jahrbuch 123 (2003) 3 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.3.339 | Generated on 2023-04-04 12:32:11 German Works Councils Old and New 341 portance of new service-sector establishments that are less likely than their manufacturing counterparts to have works councils. Although works councils are a core element of the dual system of industrial relations in Germany, representative and up-to-date information on their fre quency and coverage is sparse. Thus, there are no official statistics on works councils, and neither the union movement nor employers' associations publish pertinent information. We note parenthetically that in its draft proposals for a new Works Constitution Act, presented contemporaneously with the publica tion of the final report of the Codetermination Commission, the Federation of German Unions demanded registration of all enterprises without works coun cils in an accessible public register (see Deutscher Gewerkschaftsbund, 1998). This request did not find its way into the new law. The present paper seeks to fill this deficit in our knowledge by providing up-to-date information on works council presence and coverage by establish ment size interval and by broad sector for both westem and eastem Germany, using data from the IAB establishment panel. Probit estimation of works coun cil incidence is also offered. In addition, we document the extent of newly established works councils and inquire into the circumstances of their forma tion. The structure of the paper is as follows. Section II provides descriptive sta tistics on works councils for the year 2000. Section III investigates the deter rninants of works council presence. Section IV tackles the separate issue of newly-established councils. Abrief interpretative section concludes. 2. Works Councils in Germany: Facts and Figures for 2000 The descriptive statistics presented in this section are derived from the IAB establishment panel. Each year since 1993 (1996), this panel has surveyed several thousand establishments from all sectors of the economy in westem (eastern) Germany.3 lt is based on a stratified random sample -using strata for 16 industries and 10 size classes -from the population of all local production units with at least one employee covered by social insurance. (Due to the stra tification of the sample, the following material is based on weighted data.) Note that private-sector undertakings which do not have insured employees cannot have a works council under the law (recall the employment size thresh old of five permanent employees, noted earlier), so that from this perspective the restriction is not a problem. Nevertheless, it is the case that a small number of public-sector workplaces employing only civil servants, who are formally 3 For more information on the IAB establishment panel, see Kölling (2000). The data are confidential but not exclusive. Those interested in using the IAB data for scientific (noncommercial) research should contact the second author:([email protected]). Schmollers Jahrbuch 123 (2003) 3 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.3.339 | Generated on 2023-04-04 12:32:11 342 John T. Addison, Lutz Bellmann, Claus Schnabel, and Joachim Wagner outside the social insurance system, are excluded from the IAB panel. Infor mation on the works council status of establishments in westem and eastem Germany is available for the 1996, 1998, and 2000 waves of the panel. For present purposes, we focus on the most recent figures for 2000. Table 1 reports information on works council presence ( defined as the pro portion of establishments with a works council) and coverage (the proportion of employees working in establishments with works councils) for the German economy as a whole, and for westem and eastem Germany, by six establish ment size classes. Only establishments with five or more employees are in cluded in the sample due to the legal threshold mentioned above. On the basis of this restriction, works councils are encountered in just one in every six Ger man establishments, and this share is only slightly higher in westem than east em Germany. Works council presence tends to increase sharply with establish ment size, from about one in ten in the smallest size class with less than 21 employees to about nine in ten among the large firms with more than 500 employees. The big picture is rather similar in westem and eastem Germany. Clearly, the positive association between establishment size and works council presence produces a value for coverage that considerably exceeds that for in cidence -slightly more than one-half of all employees in Germany work in establishments that have a works council. Both works council presence and coverage varies widely by sector. Table 2 shows that nearly two-thirds of all establishments in the public sector have a works council (termed a Personalrat) compared to only one in eight establish ments in the private sector (where the works council is termed a Betriebsrat). Although these figures do not differ materially between westem and eastem Germany, we find a remarkable difference in works council incidence when we split the private sector into its manufacturing (including construction) and services components. In westem Germany, works council presence is higher in manufacturing than in services, and folklore has it that the shrinking share of the manufacturing sector combined with a growing share of private services is one of the main reasons for the long-run decline of works council presence in (westem) Germany. By contrast, in eastem Germany works council pre sence is much smaller in manufacturing than in services, while works council coverage in manufacturing is about 22 percentage points lower than in westem Germany. Table 3 combines information on establishment size class and sector. In the public sector, about one-half of the smallest units have a Personalrat, and nearly all of the larger units with more than 100 employees have the institu tion. In the private sector in both westem and eastem Germany, works council presence tends to increase with firm size in manufacturing and private ser vices. Observe that in both regions works councils are a very rare species in the smallest establishments in the manufacturing sector. Schmollers Jahrbuch 123 (2003) 3 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.3.339 | Generated on 2023-04-04 12:32:11 German Works Councils Old and New Table 1 343 Distribution and Coverage of Works Councils by Establishment Size in 2000 (in percent, establishments with five employees or more) Establishment westem Germany eastem Germany Germany size interval (number of Presence Coverage Presence Coverage Presence Coverage employees) 5 -20 9.3 10.5 7.8 9.8 9.1 10.4 21 -50 29.9 31.5 29.9 30.8 29.9 31.3 51 -100 52.9 53.4 51.2 51.3 52.6 53.0 101-200 68.6 69.5 69.1 69.7 68.7 69.5 201-500 81.4 82.5 76.2 77.4 80.6 81.8 501 and above 93.3 93.5 82.1 86.3 91.7 92.6 Average 16.6 54.1 15.4 47.1 16.3 53.0 Notes: Presence refers to the share of establishments with a works council. Coverage denotes the share of employees working in an establishment with a works council. Source: IAB Establishment Panel, Waves 8 (westem Germany) and 5 (eastem Germany), 2000. Table 2 Distribution and Coverage of Works Councils by Seetor in 2000 (in percent, establishments with five employees or more) Seetor westem Germany eastem Germany Germany Presence Coverage Presence Coverage Presence Coverage Public sector 62.9 88.9 67.4 90.2 63.5 89.2 Private sector (incl. Agriculture) 12.4 47.5 11.8 37.0 12.3 45.8 Manufacturing (incl. Construction) 13.4 58.2 8.7 36.4 12.4 54.9 Private services 11.8 40.0 14.3 39.7 12.2 40.0 All sectors 16.6 54.1 15.4 47.1 16.3 53.0 Note: See notes to Table 1. Source: IAB Establishment Panel, Waves 8 (westem Germany) and 5 (eastem Germany), 2000. Schmollers Jahrbuch 123 (2003) 3 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.3.339 | Generated on 2023-04-04 12:32:11 J 1 u.) 1 u.) Table 3: Distribution of Works Councils by Establishment Size and Seetor in 2000 (share of establishments with five employees or more that have a works council, in percent) Establishment size Western Germany eastern Germany Germany interval (number ManuPrivate Public ManuPrivate Public ManuPrivate of employees) facturing Services Seetor facturing services sector facturing services 5 -20 3.0 7.2 51.7 2.7 8.4 46.9 3.0 7.4 21 -50 27.5 22.6 75.7 14.2 31.5 78.3 24.3 24.1 51 -100 54.6 40.4 84.2 42.9 46.5 94.5 52.4 41.5 101-200 62.0 64.7 95.4 67.3 62.1 100.0 62.9 64.3 201-500 85.3 71.2 93.3 79.4 67.8 97.9 84.7 70.6 501 and above 93.9 88.0 99.5 85.3 77.8 87.3 93.3 86.5 Average 13.4 11.8 62.9 8.7 14.3 67.4 12.4 12.2 Source: IAB Establishment Panel, Waves 8 (westem Germany) and 5 (eastem Germany), 2000. Public sector 51.2 76.3 85.7 96.2 94.0 96.7 63.5 i :--3 i i � "' C/1 � ? 8. OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.3.339 | Generated on 2023-04-04 12:32:11 German Works Councils Old and New 345 Since the public sector can be regarded as a special case where works coun cils are more or less part of the organization (and bureaucracy), it may be sensible to concentrate on the private sector and on employees' decisions to establish a works council there. The broad picture regarding works council presence and coverage in the private sector of the German economy at the start of the new millennium can be summarized as follows: works councils are elected in only one of every eight establishments, although this frequency in creases with establishment size. lt is larger for manufacturing than services in westem Germany, and conversely for eastem Germany. But the positive rela tion between establishment size and works council presence means that the number of workers covered by councils is greater than the incidence of coun cils: no less than 47.5 (37) percent of all employees in westem (eastem) Ger many are working in works council establishments. 3. The Determinants of Works Council Presence Why is it that so many establishments that could have a works council under the law do not in practice have one? Even if elections of works councils are neither mandated nor automatic, in a council-free regime all that is strictly necessary to bring about an election and thence a works council is for a small group of at least three employees ( or a union that claims to have at least one member in the plant) to call for a works meeting where an electoral board will be elected which will in turn call the election, implement it, and announce the result. The requirements are hardly onerous. Given the far-reaching codetermi nation rights of the works council, the interesting question is why some estab lishments but not others are covered by the institution. The first answer to this question can be found as a byproduct of the inquiry by FitzRoy and Kraft (1987) into works councils' efficiency effects. The authors estimated a probit equation for the probability of observing a works council using a small sample of 61 / 62 firms in the metal working industry in 1977179. They found that the probability of works council presence signifi cantly increases with the number of employees and is lower in urban loca tions. Based on a much larger sample of 1,867 firms in the private sector in 1987, Frick and Sadowski (1995) presented (without further amplification) a logistic regression in which the dependent variable is a works council dummy. They reported that works councils are more likely the larger / older the firm and the more qualified its workforce. Firms experiencing greater difficulty in recruit ing qualified personnel and those with large swings in labor demand were also found to be more likely to have a works council. Conversely, works councils were less likely to be observed the higher the share of female workers and the greater the proportion of part-timers in the workforce. Schmollers Jahrbuch 123 (2003) 3 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.3.339 | Generated on 2023-04-04 12:32:11 346 John T. Addison, Lutz Bellmann, Claus Schnabel, and Joachim Wagner In a cmnprehensive study that focuses on the determinants of works council presence, Addison, Schnabel, and Wagner (1997) use data from a regional sample of manufacturing establishments in Lower Saxony, namely, the Hann over Firm Panel.4 Employing data from the first wave of the panel, adminis tered in 1994, the authors link works council presence to structural, worker composition, and participation variables. The key structural regressors are firm size (and its square), firm age, branch plant status, and a detailed set of (30) industry dummies. Our earlier descriptive material has already flagged the likely importance of establishment size. There are a number of factors that would lead us to anticipate a positive association between size and works council presence. Most important perhaps is the fact that the formal authority of the works council is increasing in establishment size, thereby providing an increased incentive for workers to elect a council in these circumstances. Gov emance considerations in larger plants might also lead both sides to embrace the institution. Less positively, the more routinized and regimented work set tings commonly attributed to larger establishments might encourage workers to demand works councils on defensive grounds. For its part, branch plant status might capture any demonstration effect that the existence of a works council in a parent firm might have on its affiliate. Older plants might be more likely to have councils because of their more established traditions of collec tive representation, and/ or different managerial ethos and type of workers. Addison, Schnabel, and Wagner argue that certain characteristics of the workforce (and the workplace) may be expected to index a taste for represen tation on the part of employees. As cases in point, they identify the propor tions of blue-collar workers, shift workers, females, and part-timers. lt is pos tulated that the two former (latter) variables are likely to be positively (nega tively) associated with works council presence for reasons that reflect both worker choice and organizational stability considerations. Finally, the authors seek to test the argument that both direct and indirect participation variables -in the form of dummies for teamworking and profit sharing, respectively -might represent a substitute in the eyes of the work force for the machinery of the works council, even if one of the functions of the works council encompasses codetermination rights in the implementation of such arrangements. They also consider profit-sharing schemes for man agers, to test one aspect of the management pressure hypothesis advanced by FitzRoy and Kraft (1987, 1990) that such schemes will encourage manage ment to seek compromise and thereby lessen the need for workers to elect works councils on defensive grounds. 4 The population of the (four-wave) Hannover panel is all manufacturing firms with at least 5 employees. The sample of establishments is stratified by firm size, with over sampling of larger plants. The first wave of the panel contains data on 1,025 establish ments. For further information on this and the remaining three waves, see Brand, Car stensen, Gerlach, and Klodt (1996). Schmollers Jahrbuch 123 (2003) 3 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.3.339 | Generated on 2023-04-04 12:32:11 German Works Councils Old and New 353 For the period 1996 / 98 information for each of the above variables is avail able for 1,950 establishments; 80 of which introduced a works council. For 1998 / 2000, the corresponding values are 2,294 and 109 plants. Ta ble 5 Probit Estimates of Works Council lntroduction Variable No works council in No works council in 1996, but one in 1998 1998, but one in 2000 Constant -1.3999** -1.0266** (-3.04) (-3.56) Establishment size 0.0037** 0.0010* (number of employees) (4.13) (2.19) Establishment size squared - 2.48 e-6* -9.06 e-8 (-2.52) (-1.14) Branch plant 0.6210** 0.5813** (dummy: 1 if yes) (4.69) (4.68) Establishment age 0.1183 [ not available] (dummy: 1 if < 6 years) (0.94) Legal status of the firm -0.3568** -0.3928** (dummy: 1 iffamily-namefirm) (-2.71) (-3.55) Covered by a collective agreement 0.2324 0.3239** (dummy: 1 if yes) (1.82) (3.21) Blue-collar workers 0.0033 0.0029 (percentage) (-1.59) (1.50) Part-time workers -0.0012 -0.0017 (percentage) (---0.42) (-0.64) Female workers -0.0029 0.0006 (percentage) (-1.18) (0.25) Shift workers 0.0031 [ not available] (percentage) (1.42) Profit situation (dummy: 0.0061 0.0571 1 if 'very good ' or 'good ') (0.05) (0.57) Eastem Germany (dummy) - 0.4155** -0.1211 (-3.23) (-1.21) lndustry dummies yes yes** n 1,950 2,294 Pseudo R2 0.150 0.129 Notes: Heteroscedastic-consistent t-values in parentheses. ** , * denote statistical significance at the 0.01 and 0.05 levels, respectively. All variables are for the first year indicated, either 1996 or 1998. Source: IAB Establishment Panel. Schmollers Jahrbuch 123 (2003) 3 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.3.339 | Generated on 2023-04-04 12:32:11 354 John T. Addison, Lutz Bellmann, Claus Schnabel, and Joachim Wagner Results of fitting the probit regressions to the data for 1996/98 and 1998/ 2000 are provided in Table 5. In both periods, there is a positive association between establishment size, branch plant status and coverage by a collective agreement, and works council formation. There is also a consistently negative relation between legal status and works council introduction. Both sets of re sult are in line with our priors. Contrary to folklore, there is no suggestion that overall business condi tions (i.e. profit situation) matter in either period. The same holds for each of the four worker composition variables, and also for the establishment age argument (included in the 1996 / 98 estimation alone for reasons of data availability). For its part, location in eastem Germany reduces the probability of works council introduction. However, this effect is statistically significant at conven tional levels only for the first period investigated. Up to this point our discussion has devolved on issues of the statistical sig nificance of the estimated coefficients and the directional influence of the cov ariates. But information on the extent of this influence -the issue of economic significance -is even more important. Although a variable that has no statisti cally significant impact can be ignored from an economic point of view, the opposite is not true. That is to say, a variable that is highly statistically signifi cant rnight have no material impact. Thus, for example, if the estimated prob ability of observing the introduction of a works council diminishes by 0.1 per cent when plant size falls from 250 to just 20 employees, we can ignore plant size in any discussion of works council introduction, irrespective of a high level of statistical significance indicated by the prob-value. Unfortunately, the estimated coefficients from a probit model cannot easily be used to make statements about the size of the ceteris paribus effect of a change in the value of an exogenous variable on the value of the endogenous variable because the size of the effect depends on both the starting level of the exogenous variable under consideration and on the values of all other vari ables in the model (see Long and Freese, 2001, 87 ff.). One way to ease inter pretation of the estimates is to construct a lirnited number of types of plant, using different values of selected variables that are both statistically signifi cant and econornically important while fixing the values of the other variables in the empirical model at their sample means in the case of continuous vari ables or at the most common frequency for dichotomous variables. To examine the economic significance of three of the variables found to be statistically significant throughout -namely, establishment size, coverage by a collective agreement, and legal status -we perform a simulation that follows this procedure. That is, we shall consider a hypothetical plant with certain characteristics, compute the probability of works council formation for this plant based on the relevant coefficient estimates (reported in Table 5), and Schmollers Jahrbuch 123 (2003) 3 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.3.339 | Generated on 2023-04-04 12:32:11 German Works Councils Old and New 355 then manipulate the value of the variables under consideration sequentially so as to illustrate their ceteris paribus impact. The simulation uses the results for 1996/98, reported in the second column of Table 5. The hypothetical plant lA has 20 employees, is a family-name business, and is not covered by collective agreement. Furthermore, this plant - and indeed all the other hypothetical establishments considered here -is not a branch plant; it is not a young plant; it has neither a 'very good' nor a 'good' profit situation; it is located in westem Germany and operates in the reference group industry; and has proportions of blue-collar, part-time, female, and shift employees that are fixed at their sample means. The estimated probability of introducing a works council for Firm lA is a tiny 0.025. Changing the legal status of the establishment firm (call this establishment type 2A) increases the probability to 0.054. Despite the very considerable increase in the likelihood of observing a new works council, it remains the case that the probability va lue is low. In plant type 3A -identical to establishment 2A in all respects other than it being covered by a collective agreement -the probability of works council formation is 0.084, much higher than that for establishment type 2A but still rather modest. To illustrate the contribution of establishment size, we compare our three types of plants (lA, 2A, and 3A) with otherwise identical establishments that have 250 (rather than 20) employees, and which we duly designate as lB, 2B, and 3B. The estimated probabilities of observing works council introduction for this group are 0.102, 0.181, and 0.248, respectively. These are considerably higher values than were obtained for the smaller comparators. As was the case for legal status and collective agreement coverage, establishment size is an important determinant of council formation. Equally, however, the magnitude of the absolute values indicate that the introduction of a works council is ex pected to be a rare event not only in small plants but also in medium-sized establishments too. 5. Concluding Remarks This paper has provided the most comprehensive information to date on the incidence and coverage of German works councils by establishment size class and by broad sector and region. Furthermore, it has offered an analysis of the determinants of works council presence, supplementing this with information on the frequency of new councils and the factors underpinning their forma tion. Our basic findings are threefold. First of all, works councils are far from ubiquitous in the firmament of smaller establishments. In 2000, only 9 percent (30 percent) of all establishments with 5 -20 employees (21-50 employees) bad a works council. The values are even lower when we single out the private Schmollers Jahrbuch 123 (2003) 3 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.3.339 | Generated on 2023-04-04 12:32:11 356 John T. Addison, Lutz Bellmann, Claus Schnabel, and Joachim Wagner sector. Given that smaller plants comprise a large fraction of the total, works councils are found in just 12.3 percent of German private sector-establish ments with 5 or more employees. Nevertheless, reflecting greater works coun cil penetration in larger establishments, a much higher proportion of German workers than German plants are covered by works councils: 45.8 percent in the private sector, and 89.2 percent in the public sector. Second of all, works council presence is not random. Most importantly, in the private sector it tends to be positively related to establishment size and coverage by a collective agreement, and it is much lower in family-name busi nesses. Finally, the introduction of a new works council is observed in less than 6 percent of all establishments in the IAB panel between 1996 (1998) and 1998 (2000). The probability of observing new councils is positively related to es tablishment size and coverage by a collective agreement, and it is again much lower in family-name undertakings. Given the scarcity of detailed, representative, and contemporary informa tion on works council frequency, the descriptive evidence provided here has intrinsic worth. Furthermore, it is relevant for at least two other reasons. Thus, the nonrandom nature of works council presence / formation means that poten tially serious biases characterize point estimates of the work council effect on firm performance derived from OLS estimation. With a few exceptions -the two most recent examples being the very different methodological approaches of Hübler and Jirjahn (2001) and Addison, Bellmann, Schnabel, and Wagner (2003) -the literature has tended to treat works councils as exogenous. More over, the rare events nature of works council introduction in hitherto uncov ered plants, and the core role of establishment size in this regard, make it im probable that the central aim of the new Works Constitution Act will be met. That aim is to shrink the extent of the codetermination free zone, which is largely made up of smaller establishments. Further, given the important role of bargaining coverage in determining works council presence, the steadily fall ing number of establishments covered by collective agreements reported by Kohaut and Schnabel (2003) is scarcely conducive to an increase in the inci dence and coverage of works councils and shrinkage of this works council free zone. Time will tel1 whether or not the legislative changes and the intensive PR campaign of the union movement will lead to a large number of new works councils in small and medium-sized establishments. Based on our findings, we doubt this will happen. But we plan carefully to monitor such develop ments -and the economic effects of new councils -in future work. Schmollers Jahrbuch 123 (2003) 3 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.3.339 | Generated on 2023-04-04 12:32:11 German Works Councils Old and New References 357 Addison, J. T. / Bellmann, L. / Schnabel, C. / Wa gner, J. (2003), The Reform of the Ger man Works Constitution Act: A Critical Assesment, Industrial Relations (forthcom ing). Addison, J. T. / Schnabel, C. / Wa gner, J. (1997), On the Determinants of Mandatory Works Concils in Germany, Industrial Relations 36, 419 -445. Brand, R. / Carstensen, V. / Gerlach, K. / Klodt, Th. (1996), The Hannover Firm Panel, Discussion Paper No. 2, Universität Hannover. Deutscher Gewerkschaftsbund [Bundesvorstand] (1998), Novellierungsvorschläge des DGB zum Betriebsverfassungsgesetz 1972, Düsseldorf: Deutscher Gewerkschafts bund, Februar. Dilger, A. (2002), Payment Schemes, Returns and Works Councils, Paper presented at the Workshop "Mitbestimmung und Mitarbeiterbeteiligung: komplementär oder sub stitutiv?", University Witten/Herdecke, February 20, 2002. Fi tzRoy, F. R. / Kraft, K. (1987), Efficiency and Internal Organization: Works Councils in West German Firms, Economica 54, 493 -504. - (1990), Innovation, Rent Sharing, and the Organization of Labour in the Federal Re public of Germany, Small Business Economics 2, 95 -103. Frick, B. I Sadowski, D. (1995), Works Councils, Unions, and Firm Performance, in: F. Buttler et al. (Eds.), Institutional Frameworks and Labor Market Performance - Comparative Views on the U.S. and German Economies, London and New York: Routledge, 46 -81. Hübler, 0. / Ji rjahn, U. (2001), Works Councils and Collective Bargaining in Germany: The Impact on Productivity and Wages, Discussion Paper No. 332, Institute for the Study of Labor (IZA), Bonn, July. Kölling, A. (2000), The IAB Establishment Panel, Schmollers Jahrbuch/ Journal of Ap plied Social Science Studies 120, 291-300. Kohaut, S. / Schnabel, C. (2003), Tarifverträge -nein danke!? Ausmaß und Einflussfak toren der Tarifbindung westund ostdeutscher Betriebe, Jahrbücher für Nationalöko nomie und Statistik 223, 312-331. Kommission Mitbestimmung (1998), Mitbestimmung und neue Unternehmenskulturen -Bilanz und Perspektiven. Gütersloh: Verlag Bertelsmann Stiftung. Lang, J. S./ Freese, J. (2001), Regression Models for Categorical Dependent Variables using Stata, College Station, TX: Stata Press. Winship, Ch./ Radbill, L. (1994), Sampling Weights and Regression Analysis, Sociolo gical Methods & Research 23, 230-257. Schmollers Jahrbuch 123 (2003) 3 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.3.339 | Generated on 2023-04-04 12:32:11 358 John T. Addison, Lutz Bellmann, Claus Schnabel, and Joachim Wagner Appendix Table Probit Estimates of Works Council Presence Variable Germany westem eastem Germany Germany Constant -1.0624** 0.1121 0.3401 (-7.75) (0.56) (1.32) Establishment size 0.0032** 0.0039** 0.0032** (number of employees) (11.79) (7.50) (8.78) Establishment size squared -7 .71e -08** -9.23e-08** -6 .14e-07** (-11.78) (-7.50) (-8.52) Branch plant 0.6030** 0.7230** 0.4359** (dummy: 1 ifyes) (14.51) (13.38) (6.35) Establishment age -0.2770** - 0.1519* -0.3871** (dummy: 1 if < 6 years) (--4.89) (-2.14) (-3.91) Legal form of the firm ( dummy: -0 .6905** -0 .5425** -1.0243** 1 iffamily-ownedfirm) (-17.37) (-11.33) (-11.75) Foreign ownership 0.2617** 0.1053 0.6626** (dummy: 1 ifforeign ownership) (3.49) (1.19) (4.22) Covered by a collective agree0.9758** 0.8707** 1.0576** ment (dummy: 1 ifyes) (26.66) (17.43) (18.57) Blue-collar workers -0.0060** - 0.0066** -0 .0053** (percentage) (-9.50) (-7.96) (-5.45) Part-time workers -0 .0033** -0.0011 -0.0062** (percentage) (-3.52) (0.85) (--4.37) Female workers -0.0021 ** -0 .0049** 0.0021 (percentage) (-2.69) (--4.73) (1.73) Profit sharing scheme for 0.1100* 0.0173 0.3703** employees (dummy: 1 ifyes) (2.12) (0.28) (3.70) Employee share ownership 0.3005** 0.3555** 0.1302 scheme (dummy: 1 ifyes) (3.67) (3.45) (0.90) Eastem Germany ( dummy) - 0.1201 ** - - (3.64) Industry dummies yes** yes** yes** n 10,515 6,498 4,017 Pseudo R2 0.4206 0.4281 0.4388 Notes: Heteroscedastic-consistent t-values in parentheses. **, * denote statistical significance at the 0.01 and 0.05 levels, respectively. Source: IAB Establishment Panel, Waves 8 (western Germany) and 5 (eastern Germany), 2000. Schmollers Jahrbuch 123 (2003) 3 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.3.339 | Generated on 2023-04-04 12:32:11