¡Descarga amosse socio productive y más Apuntes en PDF de Relaciones Laborales y Recursos Humanos solo en Docsity!
Industrial and Labor Relations Review , Vol. 64, No. 4 (July 2011). © by Cornell university. 0019-7939/00/6404 $05.
- Thomas Amossé is a researcher of the Centre d’etudes de l’emploi (Cee) in Paris. He is also a research fellow with the maison des Sciences de l’Homme (mSH) Ange Guépin in Nantes. Thomas Coutrot is an economist at DAreS (Direction de l’animation de la recherche, des études et des statistiques) at the French ministry of la- bour. To obtain results not listed in the paper, interested readers may write to the authors at thomas.amosse @cee-recherche.fr or [email protected].
SoCIo-ProDuCTIVe moDelS IN FrANCe: AN
emPIrICAl DYNAmIC oVerVIeW, 1992–
THomAS AmoSSe AND THomAS CouTroT*
The authors assess the evolution of French workplaces over the periods 1992–1993, 1998–1999, and 2004–2005 within the framework of socio-- productive organizational models. using French employment relations sur- vey data (rePoNSe), they analyze the empirical relevance of these models in light of three dimensions: work organization, human resource management (Hrm), and industrial relations. results indicate the emergence of two “in- novative” models: the “Toyotist” model, featuring just-in-time organizational methods, innovative Hrm systems, high employee participation, and low union presence; and the “neo-Taylorist” model, featuring low innovation, low investment in Hrm, strict managerial control with minimal employee partici- pation, and average union presence. Data also reveal the existence of an ad- ditional model: the “public-in-transition” model, featuring a mixed Hrm focus that is part individual and part collective; moreover, workplaces feature a high and varied union presence and autonomous work groups. overall findings indicate a dual movement of socio-productive models: the development of the Toyotist model occurred mainly between 1992 and 1998 for some workplaces whereas for others, such as those aging or becoming privatized, there was a shift to the neo-Taylorist model for the period 1998–2004.
I
n the face of continual competition in product, labor, and capital markets, firms are constantly adapting their organizational methods. In order to attract clients, recruit staff, and attract investors, they must imple- ment marketing, human resource manage- ment (Hrm), and financial strategies. There is a consensus among researchers that the strategies adopted by firms in these three do- mains must be both mutually reinforcing (internal fit) and consistent with the firm’s
socio-economic environment (external fit) (Dyer and Kochan 1994). They do not agree, however, on the relative importance ac- corded to internal or external fit. much of the literature on new methods of work and firm organization therefore con- centrates on high performance work organi- zation (HPWo) (osterman 1994), or the so-called “dominant model theory” outlined by Dyer and Kochan (1994). According to this theory, the adoption of “high road” prac- tices, featuring autonomous teams, multi- skilling, continuous training, and employee incentives, would enable a firm to improve its economic performance in any context (Department of labor 1994) and thus be in- trinsically more effective than those adopt- ing the “low road” strategy, characterized by low wages, a Taylorist work model, and short-term contracts. In contrast, a theory insisting on the importance of external fit
SoCIo-ProDuCTIVe moDelS IN FrANCe 787
predicts that optimal practices per se do not exist independently of a firm’s environ- mental context and strategic choices. The “multiple model theory” (Dyer and Kochan 1994; Appelbaum and batt 1994) acknowl- edges the diversity of efficient organizational methods, and the issue becomes one of com- plementarities (milgrom and roberts 1990; Aoki 1994; Amable 2005) between combined strategies and practices adopted in different domains of production. For example, an in- novative practice (such as increasing em- ployees’ autonomy) can reinforce a given strategy’s effectiveness (such as permanent innovation) but may be unsuitable in a dif- ferent context (such as when pursuing a low- cost strategy). In this paper, we use a multiple model approach, based on the hypothesis that a va- riety of effective socio-productive models enable firms to deal with the uncertainties with which they must cope at both organiza- tional and market levels. After reviewing the existing literature concerning the different socio-productive models, we use data from the rePoNSe survey conducted in France among 3,000 workplaces during the periods 1992–1993, 1998–1999, and 2004–2005 to evaluate the empirical pertinence of these theoretical models, taking into account three dimensions: work organization, Hrm, and industrial relations. The longitudinal nature of our data enables us to provide a dy- namic picture of firms’ strategies over time.
socio-Productive Models: A Brief Review of Literature
In the early 1960s, management research- ers developed the contingency theory (burns and Stalker 1961), which posited the need for coherence between a firm’s internal or- ganization and the nature of its external markets—between structure and environ- ment. Specifically, a formalized or vertical structure necessarily adapts to a stable envi- ronment whereas a changing, innovative environment requires a more flexible, horizontal organization. At the end of the 1980s, this type of approach was taken up by economists (Aoki 1988; milgrom and roberts 1994). Theorists of the “J” firm, such
as Aoki, added social regulation to the model. The “J” firm is characterized by its competitive strategy (differentiation within a mass market), its organizational model (horizontal communication and workers’ autonomy), and its industrial relations sys- tem (the labor union, representing perma- nent staff, facilitates employee cooperation and pushes management to share productiv- ity increases equitably between employees and shareholders).
From “Factory Regime” to “socio-Productive Models”
During the 1980s, American radical econ- omists had already developed a systemic ap- proach to the firm, based on the marxian concept of the “factory regime” (burawoy 1983). A regime is defined as an institutional mechanism situated “between the labor pro- cess and the State” that extends beyond the firm’s boundary but accounts for the coher- ence between the nature of competition within the product market, labor organiza- tion methods, and labor-power reproduc- tion ( ibid. : 603). burawoy distinguished between two types of regimes: “despotic” ones in which a dominant employer perma- nently subjects employees to the threats of unemployment and job insecurity, and “he- gemonic” ones in which employees volun- tarily consent to their own exploitation because they have resources (qualifications, trade-unions, labor law, social protection) that endow them with the bargaining power to negotiate the terms of this exploita- tion with their employers. With the global- ization of capital, burawoy diagnosed the emergence of an additional “despotic- hegemonic” regime that is based both on “the arbitrary tyranny of the overseer aimed at individual workers” and “the ‘rational’ tyranny of capital mobility over the collective worker” ( ibid. ), with permanent threats of closure and overseas relocation of entire workplaces. returning to the firm level, three types of radical uncertainty threaten its survival: un- certainty related to market competition, un- certainty due to the complexity and fragility of in-house technical and organizational
SoCIo-ProDuCTIVe moDelS IN FrANCe 789
tion of day-to-day decision-making, horizon- tal coordination methods rather than vertical hierarchies, and a high degree of employee cooperation. In terms of industrial relations, both imply direct employee participation mechanisms (quality circles, semi-self-managed groups, project teams, and so on); a “pacified” social climate achieved by highly individualized manage- ment-employee relations; and high internal and external employee mobility. moreover, both enable a certain fragmentation of work communities. beyond these similarities, the knowledge-creating firm, in principle, dis- tinguishes itself from the lean production model in three distinct ways—employees possess higher skills; they are allowed greater discretion at work; and they are given more complex and less repetitive tasks (lorenz and Valeyre 2005).
Overlapping Models Observed throughout the economy
empirical observations typically describe the juxtaposition of innovative and tradi- tional models (Appelbaum and batt 1994); however, traditional models have been modernized through the introduction of new technologies. Neo-Taylorism or neo- Fordism thus combines the traditional char- acteristics of Taylorism and Fordism with the electronic surveillance of employee perfor- mance and computer-assisted methods of production and design. The “neo-liberal networked firm” (Coutrot 2002) stems from a hierarchical integration of establishments or firms coming from different models: knowledge-creating at the top of the net- work, neo-Fordist or “lean production” at a first subsidiary or outsourcer level, neo- Taylorist at a second level, and so on. This articulation is designed to attain sharehold- ers’ profitability standards at the top end of the network (Palpacuer, Seignour, and Vercher 2006). Furthermore, the autono- mous dynamics of industrial relations, nota- bly at both the regional and national levels, leave their footprint on firms’ indus- trial relations models. Traditions, histories, and institutions contribute in shaping how interests and conflicts are represented in
ways that, to a certain extent, escape eco- nomic determinism or strategic intentional- ity (Poutsma, ligthart, and Veersma 2006). The concept of the socio-productive model is thus complex, standing at the cross- roads of various social sciences: industrial and political economy, management, and sociology (of work, organizations, and col- lective action). A country’s industrial special- ization, its economic performance, and its social cohesion depend heavily on the predominant socio-productive model and, more globally, the institutional (lack of) complementarities at the macroeconomic level (Aoki 1994; Amable 2005; lung 2005). In what follows, we attempt to answer em- pirically two specific questions. First, are the theoretical models proposed in the litera- ture consistent with the configurations ob- served de facto in the French economy? Second, over the fifteen years covered by the rePoNSe surveys upon which we rely, is it possible to detect the decay of traditional models and the rise of one or more innova- tive models?
socio-Productive Models Viewed through the Lenses of statistical surveys: Questions of Data and Methods
For our analyses, we used different waves of the rePoNSe survey (1992–1993, 1998– 1999, and 2004–2005) as well as different datasets from within those surveys (manage- ment and employee questionnaires).
Data: the French RePONse surveys
The rePoNSe survey was largely inspired by its british counterpart, WIrS ( the Work- place Industrial Relations Surveys) , conducted in 1980, 1984 and 1990, and later becoming WerS ( Workplace Employment Relations Sur- veys) in 1998 and 2004 (see blanchflower et al. 2007 for a recent synthesis of what has been learned through the WI(e)rS series). like the british version, the rePoNSe sur- vey is conducted under the aegis of the min- istry of labor and collects the opinion not only of a senior manager in each workplace
790 INDuSTrIAl AND lAbor relATIoNS reVIeW
but also of its employee representatives (if they are present) as well as of a sample of employees. Also like the british survey, rePoNSe interrogates employers and rep- resentatives by means of face-to-face inter- views and employees via a self-administered questionnaire. The questions cover a wide spectrum of themes ranging from the struc- ture of ownership to the conditions and or- ganization of work, to industrial relations and Hrm practices. The survey also includes a panel of workplaces that are followed from one survey wave to the next. (For a detailed account of the rePoNSe survey, including its history and methodology, see Amossé and Coutrot 2008). An overview of the different samples, particularly the ones we use here, is presented in Appendix Table 1. In our study, we used primarily a pooled sample (n 5 6,265), which aggregates three samples of 1,744 (from rePoNSe 1992– 1993); 2,256 (from 1998–1999); and 2, workplaces (from 2004–2005). each of these samples is representative (when weighted) 1 of all French workplaces with 50 employees or more in the competitive sector (exclud- ing agriculture) at the time of the survey. We also used a sample of 6,128 employees working in establishments included in the 2004–2005 sample. Finally, in order to pres- ent a dynamic perspective, we drew on two panels. The first includes 371 workplaces, surveyed in 1992–1993 and in 1998–1999; the second includes 742 workplaces, surveyed in 1998–1999 and in 2004–2005. To maintain consistency with the theoreti- cal orientation sketched above (which in fact framed the survey from the outset), we built a set of indicators for each dimension of the socio-productive models, from work organi- zation to industrial relations, including Hrm (Appendix 1A). The choice of indica- tors was limited by the availability of the same variables throughout the three editions of the survey. For example, we could not in- clude variables on competitive strategy in our analysis of the pooled sample, since
(^1) The weights correct for unequal inclusion probabili- ties due to the sampling design and include post- stratification by size and industry.
employees were asked about this differently in the 1992–1993 and following editions. Nevertheless, the survey does enable us to succinctly describe how organizational prac- tices, labor management methods, and in- dustrial relations have evolved in workplaces of more than 50 employees in the competi- tive sector, between the period 1992– (see Table 1).
Main trends Observed through Univariate Analyses
over the 14 years that our study covers, ISo (International organization for Stan- dardization) quality norms continue to expand under the influence of prime con- tractors (Gorgeu, mathieu, and Pialoux 1998), but computer-assisted production technologies have reached their full devel- opment; organizational changes involving the remodeling of company boundaries (such as “focus on core business” and “out- sourcing”), however, have become less fre- quent. After a period of intensive managerial innovation in the 1990s as well as the reduc- tion of working hours implemented by the left-wing government between 1998 and 2002, organizational change seems to have slowed down (bué, et al. 2007; Greenan, Guillemot, and Kosoglu 2009). The develop- ment of participatory mechanisms such as quality circles, regular meetings, and “ex- pression groups” reflects management’s long-term strategy to increase employee in- volvement and to facilitate the implementa- tion of profound organizational changes. Concerning Hrm, we observe a signifi- cant progression in the use of individual performance appraisal and assessment inter- views, notably for non-executives (executives were already subject to these practices at the beginning of the period under review), as well as in flexible wage policies. Annual col- lective wage increases tended to be replaced by a combination of collective increases and flexible or individualized increases based on performance. New job classification meth- ods were less extensively introduced between 1996 and 1998; perhaps the depressed eco- nomic climate of previous years dissuaded management from engaging such a “formal
792 INDuSTrIAl AND lAbor relATIoNS reVIeW
consolidated while on the other individual- izing Hrm practices aimed at reinforcing direct relations between management and employees rose. However, we see an increase in trade-union presence and increases in individual and collective conflicts. In the French case, and even at this very first step of the analysis, a discrepancy between the pre- dictions of certain theoretical models and the empirical evolution is evident; far from creating peaceful industrial relations, the in- troduction of organizational and managerial innovations seems to have coexisted with fairly high levels of conflict between 1992 and 2004.
Multivariate Methods Used to empirically Identify and characterize socio-Productive Models
In order to better understand the corre- spondence between the practices and strate- gies among the different dimensions of socio-productive models, we first imple- mented a multiple Correspondence Analysis (mCA) with the Ward metric 3 on the pooled sample with, as active variables, the categori- cal indicators relating to work organization, Hrm, and industrial relations at the work- place. 4 We then used a Hierarchical Cluster Analysis (HCA) based on the first axes pro- vided by the mCA. We first describe the char- acteristics of workplaces and workers in each class, then relate the classes to the socio- productive models. Greenacre and Hastie (1987) noted that mCA is the equivalent of principal compo- nents analysis for categorical data. As in the usual principal component analysis, the successive axes may be regarded as optimal in repeated scaling programs. With the
(^3) For a comprehensive description of this method, com- putational details, and its applications, see Greenacre (1984). We have used SAS’s “ proc corresp” with the “ mca ” option. (^4) As noted above, the competitive strategy should theo- retically be included in an analysis of the active varia- bles, but the relevant indicators are unavailable in the 1992–1993 survey and the questions concerning this domain changed between 1998–1999 and 2004–2005; we include them as supplementary variables for the last survey (see Table 3).
Ward metric, any set of scale values for all of the categorical indicators used in the mCA imply a score for each respondent (here each workplace), in which the score is the average scale value of the categories into which the workplace falls. The first principal axis is then defined so that the scale value provided by the positions on it yield scores with maximum variance. In other words, principal axes are inertia maximizing, in which inertia is the weighted average of squared (chi-squared) distances from the centroid to the projections of the category points on them. main displays of mCA are graphical representations of principal axes, which are optimal scale values for the cate- gorical indicators used in the analysis. For instance, Figure 1 displays the configuration of category points that maximize the disper- sion of the workplaces in a planar display. Furthermore, each principal inertia can be decomposed into individual components ac- cording to each indicator or category. The study of these components, or “contribu- tions to inertia” (see Appendix Table 2), is an important feature of the geometric inter- pretation of the analysis. The categories that contribute highly to a principal axis (these are in bold type in Figure 1) have, in effect, largely determined the orientation and thus the identity of the corresponding axis. The mCA is interesting not only because it provides a graphical representation of the structural patterns of the workplace popula- tion in terms of work organization, Hrm, and industrial relations, but also because it aids in summarizing the abundance of infor- mation provided by 17 categorical indicators (corresponding to 48 categories) observed for 6,265 workplaces. Indeed, in the cluster analysis, we did not use all of the categorical indicators initially taken into account in the mCA or—the exact equivalent—all of the principal axes it provides. We rely only on the first six principal axes, which correspond to about one-third of the total inertia, in order to avoid having some classes defined by only a few workplaces with very specific profiles. Cluster analysis, and notably HCA, is widely used to classify multivariate data into sub-groups. We do not offer much de- tail on the method here; using the principal
SoCIo-ProDuCTIVe moDelS IN FrANCe 793
Figure 1
. The Four Socio-Productive
m
odels and the Categorical Indicators
u
sed to Define Them, as Displayed in the First Principal Plan
Notes:
The four classes are displayed as supplementary variables, whereas all the 48 categories of the 17 indicators are active variables in
the
m
CA. Categories of indicators that contribute the most to the two first axes are in bold type (some of them contribute both to the first
and to the second axis; see Appendix Table 2 for their detailed contributions. For more details on their labels, see Appendix Table 2. Field:
Workplaces with 50 employees or more in the non-agricultural trading sector.
Source:
“m
anagement Questionnaire”, pooled sample, (n = 6,265), 1992–1993, 1998–1999, 2004–
re
Po
NS
e^
surveys, Dares.
SoCIo-ProDuCTIVe moDelS IN FrANCe 795
In order to constitute a typology by an HCA, we retained the workplaces’ coordi- nates on these four axes, to which we added the fifth and sixth. The four empirical classes emerging from the cluster analysis represent workplace profiles, which combine work or- ganization methods, Hrm, and social regu- lation practices (insofar as they were highlighted by the rePoNSe survey) in a specific manner. To compare these classes
with the socio-productive models as they are outlined in the theoretical literature is no more than a heuristic tool, given that they depend largely on the arbitrary nature of methodological choices. In effect, the num- ber of classes is set according to the elbow criterion, but this same criterion applied to other specifications (e.g., when retaining eight principal axes and not six) supplies six classes. Nevertheless, the first two principal
Table 2. Characteristics of the Four Socio-Productive model Profiles (Active Variables*) (percentage of workplaces)
Simple Control (n 5 509)
Neo-Taylorist (n 5 2 230)
Toyotist (n 5 1565)
Public-in- Transition (n 5 1565)
Work organization ISo standard **9**** 26 49 12 Computer-assisted production, robot (^8) 27 20 ( 2 ) 8 At least two organizational innovations (refocusing, outsourcing, streamlining) 15 17 19 32 At least two participatory devices (quality circles, employee meetings, expression groups) 43 44 69 48 ( 2 ) Human Resources Management Individual appraisal interviews (for all employees) 32 ( 2 ) 22 66 55 Profit sharing agreement 25 44 59 55 No general wage increases (either for management or non-management) 13 13 33 12 recent change in job classifications 45 29 32 42 Industrial Relations Intensive communication with employees (newsletter, mailings.. .) 32 ( 1 ) 21 47 60 Participatory devices (ideas box or open-door day) (^29) 27 49 31 ( 2 ) Firm is member of an employer federation 61 ( 1 ) 60 ( 2 ) 63 61 ( 2 ) Participates in employer network (industry federation and local employer association) 37 ( 1 ) 31 ( 2 ) 40 45 Presence of union(s) representative(s) 4 62 51 ( 2 ) 85 Presence of at least two elected delegates (employee representatives, work council, HSWCC) 25 89 85 95 Problem of absenteeism (according to by management) 39 ( 2 ) 67 29 47 Numerous sanctions against employees 34 35 13 11 Collective action (strike during the last three years) 3 25 10 53 size of the different classes 13 36 33 18
Notes: *See Appendix 1A for a precise definition of the indicators used to define the classes. ** All figures are weighted. The regression models also include indicators of workplace size (50–99; 100–199; 200–499; 500 and more) and sector (15 categories roughly corresponding to first level of the NAICS). “ 1 ” and “ 2 ” signs indicate the direction of the net effect (all being equal) when it is not clear. The detailed results of the models are presented in Appendix Table 3a. The figures in bold type identify those items which, in a logistic regression model explaining whether a workplace belongs to the socio-productive model noted in the column heading, are statistically significant at the .05 level. Field: Workplaces with 50 employees or more in the non-agricultural trading sector. Source: “management Questionnaire,” pooled sample, (n 5 6,265), 1992–1993, 1998–1999, 2004–2005 rePoNSe surveys, Dares.
796 INDuSTrIAl AND lAbor relATIoNS reVIeW
Table 3. Characteristics of the Four Socio-Productive models (Additional management Variables* in 2004–2005) (percentage of workplaces)
Simple Control (n 5 130)
Neo–Taylorist (n 5 787)
Toyotist (n 5 756)
Public-in- Transition (n 5 592)
Legal and ownership structure single site firm 50 44 38 33 Workplace older than 20 years 55 67** 65 74 Family owned 36 37 29 14 under public control (State, local authorities) 5 2 3 13 belongs to a larger corporation 32 54 61 57 listed firm 16 30 42 37 economic strategy and positioning market spread Local (45) Regional (20) Worldwide (30) local (31) market share over 25% 27 34 25 30 Predictable market 31 26 29 41 Stable market 57 55 66 67 main element of economic strategy Service quality (42), price (22)
Price (24) Product quality (21), Innovation (10)
Sevice quality (42), no strategy (7) Substantial market power 25 20 27 18 main benchmark in price fixing Costs (30) Costs (27) market (37) regulation (20) Firm’s primary objective Quality (30) Profitability (29) Profitability (37) budget (28) Prime contractor 44 62 63 63 Sub-contractor (for at least 10% of turnover) 17 20 19 16 Non-profit organisation 7 8 5 15 economic health High profitability (according to management) 25 22 30 22 Growth in business activity 63 52 55 51 Growth in total number of employees 50 41 46 39 Growth in number of executives (“cadres”) 25 28 41 32 Growth in number of white collar employees (“employés”) 41 31 30 22 Growth in number of blue collar employees (“ouvriers”) 23 26 20 17 Human resource management and work organization Training budget amounts to 2% of pay roll 44 48 66 64 Just-in-time arrangement with customers 33 42 42 33 Integrated management solutions (erP) 26 33 53 39 Significant technological change (in last 3 years) 13 11 17 22 Significant organizational change (in last 3 years) 27 31 34 43 Product innovation (in last 3 years) 35 30 41 45 Work is strictly prescribed 62 75 58 65 low employee autonomy 46 52 37 41 continued
798 INDuSTrIAl AND lAbor relATIoNS reVIeW
communication policies, and individual ap- praisal interviews. Another commonalty these workplaces share is the fairly intensive use of sanctions against employees. At the same time, and in contrast to the simple con- trol profiles, these workplaces are much more high-tech, using many Information Communication Technology (ICT) devices, and are plagued with high rates of absentee- ism, which may indicate that employees are dissatisfied with working conditions or that they must perform difficult jobs. being on average larger, these workplaces more fre- quently have a profit sharing scheme but less
frequently have adopted a new job classifica- tion scheme, confirming their relatively low investment in Hrm. Participation in em- ployer networks is low and union presence is average. The supplementary variables confirm that this profile closely resembles the neo- Taylorist production model, which features just-in-time organization; well developed (employee) multi-skilling; meticulously pre- scribed work; low latitude for employees to intervene in work-related incidents (like production breakdowns); competition based on low prices; and a lower profit margin
Table 4. Characteristics of the Four Socio-Productive models (Additional employee Variables* in 2004–2005) (percentage of employees)
Simple Control (n 5 295)
Neo-Taylorist (n 5 2 071)
Toyotist (n 5 2 121)
Public-in- Transition (n 5 1 641)
Work organization Always or often works in a hurry 32 27 27 27 Chooses one’s work method 75* 82 87 87 Personally handles incidents 51 51 57 57 Participates in work related meetings 53 58 ( 2 ) 73 74 Has individual appraisal interviews with superiors 37 44 67 68 Reasons for investing in one’s work To gain respect from peers 27 23 28 29 To satisfy clients, users 76 68 73 76 Fear of losing one’s job 19 16 12 11 Obstacles to investing in one’s work Working conditions 19 19 15 16 lack of recognition 42 40 33 36 Industrial relations Has participated in meetings with employee representatives 19 29 25 ( 2 ) 35 Has participated in a work stoppage 3 23 14 ( 2 ) 28 Has participated in another form of collective action (petition, rally etc.) 7 21 17 ( 2 ) 28 Is a union member 3 8 ( 2 ) 6 13 Indicates that “management consults employees in the event of tensions within the workplace” 30 44 41 44
Notes: *Questions used to define the indicators can be found on http://www.travail-solidarite.gouv.fr/ImG/pdf/ QueST_Sal_2004–2005.pdf. For more details, please email the authors. ** All figures are weighted. The regression models also include indicators of workplace size (50–99; 100–199; 200–499; 500 and more) and sector (15 categories roughly corresponding to first level of the NAICS). “ 1 ” and “ 2 ” signs indicate the direction of the net effect (all else being equal) when it is not clear. The detailed results of the models are presented in Appendix Table 3c. The figures in bold type identify those items which, in a logistic regression model explaining whether a workplace belongs to the socio-productive model noted in the column heading, are statistically significant at the .05 level. Field: Workplaces with 50 employees or more in the non-agricultural trading sector. Source: “employee Questionnaire” (n 5 6,128), 2004–2005 rePoNSe survey, Dares.
SoCIo-ProDuCTIVe moDelS IN FrANCe 799
than most competitors. These workplaces are almost never public utilities, but they are frequently family-owned firms. The related industries include not only manufacturing (in particular intermediary and production goods) but also road transportation, retail trade, and services cleaning. This model characterizes 36% of workplaces and 35% of employees. The employees queried in these work- places largely corroborate the analysis result- ing from their managers’ responses—they have less autonomy, insofar as they frequently admit that they must call for assistance in case of an incident at work. moreover, em- ployees in these workplaces more frequently complain about “poor working conditions” as being an obstacle to them making a greater investment in their work. 9 At the same time, they do not adopt the exit strate- gies (Hirschman 1970) that their managers indicate. That is, a high proportion of em- ployees (almost one-half if one considers both work stoppages and other forms of col- lective action) has participated in a collec- tive dispute during the last three years preceding the survey, which in practice tends to attenuate this model’s despotic nature.
toyotist
The third workplace profile (33% of workplaces and 34% of employees) is radi- cally different from the two preceding mod- els. The Toyotist model features widespread use of innovative Hrm systems, whether in terms of communication devices, employee participation, individual appraisal interviews, profit-sharing agreements, or wage flexibil- ity and ISo standards. These workplaces have exceptionally peaceful industrial rela- tions. managers report low absenteeism and few individual sanctions or collective work stoppages. union presence is somewhat lower than in other workplaces of equivalent size. According to the supplementary variables, work organization methods confirm that this profile is close to both the lean production
(^9) The difference between these and employees at work- places in other models is not statistically significant.
and knowledge-creating models: work is var- ied, management declares that employees benefit from a high degree of autonomy in handling work-related incidents, and train- ing budgets are high. Just-in-time organiza- tional methods and integrated management solutions are common, indicating the pres- ence of horizontal coordination systems. These companies belong to major firms listed on the stock exchange, operate in in- ternational markets, and are often regarded as stable; moreover, their competitive strate- gies are based on innovation and product quality, and profitability is considered rela- tively high compared with competitors. They are typically found in the consultancy sector, IT, high value-added manufacturing indus- tries (automotive and luxury goods), and specialized wholesale and retail trade. In fact, the work organization indicators we use here are not precise enough to allow for a clear empirical distinction between the lean production and knowledge-creating models, which is why we qualify this profile as “Toyotist,” a somewhat less precise label, in reference to Aoki’s (1988) model, even if French unions do not have the same close relationship with their firm than their Japanese counterparts. In these workplaces, employees’ state- ments confirm those of their managers.’ To wit, compared to employees of other mod- els’ workplaces, they more often declare that their work is not tightly prescribed; that they have an annual individual appraisal inter- view; that they invest in their work to “gain recognition from colleagues”; that they may potentially earn a promotion or a wage increase;^10 and that they rarely participate in collective conflicts. managerial hegemony appears to be salient in this socio-productive model.
Public-in-transition
The fourth configuration is not predicted by any of the theoretical models discussed above. These workplaces practice organiza- tional, technological, and product innova-
(^10) Please see footnote 9.
SoCIo-ProDuCTIVe moDelS IN FrANCe 801
utilities continues to shape the productive base and especially industrial relations, de- spite the active privatization policies carried out since the mid 1990s.
the Dynamics of socio-Productive Model
From one edition of the survey to the next, the relative importance of the differ- ent socio-productive models present in the French economy changes under the impact of two mechanisms—workplace demograph- ics and transitions between models. First, some establishments disappear, others are created, and the resulting impact modifies the relative weight of the different models. Second, workplaces can switch between models; panel analysis reveals these transi- tions between models. using the panel surveys—one comprising establishments queried in both 1998– and 1998–99, and another comprising those present in both 1998–1999 and 2004– 2005—we can describe how workplaces shifted from one model to another (Tables 6 and 7). Panel data allow us to compute the flows of workplaces and employees that changed classes between the two surveys as
well as determine how the relative weight of each class has increased or diminished (see Figures 2 and 3). The two results will be com- pared with the cross-sectional data on change presented in Table 5.
Declining Models: “Public-in- transition” and “simple control”
between 1998 and 2004, both the public- in-transition and simple control models de- clined whereas the Toyotist model became more widespread. The prevalence of neo- Taylorist workplaces remained relatively sta- ble. The decline of the simple control model must not be overstated, however, consider- ing that in this study we ignore the changes that took place in workplaces with fewer than 50 employees. Concerning the public and Toyotist models, their relative importance was inverted over a period of less than fif- teen years: in 2004–2005, they represented 25% and 34% of employees respectively, whereas that percentage was 30% and 24% in 1998–1999. During the most recent survey, the Toyotist and neo-Taylorist work- places account for equivalent numbers of employees, well above public-in-transition workplaces.
Table 5. The Dynamics of Socio-Productive models (cross-sectional data)
1992–1993 1998–1999 2004–
Percentage of workplaces* Simple control 16 12 11 Neo-Taylorist 38 34 36 Toyotist 25 36 35 Public-in-Transition 21 16 17 total 100 100 100
Percentage of employees* Simple control 11 5 6 Neo-Taylorist 35 34 35 Toyotist 24 36 34 Public-in-Transition 30 25 25 total 100 100 100
Notes: * Figures are weighted with two different weighting sets: that used in the upper panel allows the analyst to infer proportions in the workplaces population; that used in the lower panel allows the analyst to infer the proportions in the corresponding employee population. Field: Workplaces with 50 employees or more in the non-agricultural trading sector. Source: “management Questionnaire,” pooled sample, (n 5 6,265), 1992–1993, 1998–1999, 2004–2005 rePoNSe surveys, Dares.
802 INDuSTrIAl AND lAbor relATIoNS reVIeW
Table 6. The Dynamics of Socio-Productive models (1992–1993 and 1998–1999 Panels)
Percentage of Workplaces
1998–
Simple Control Neo-Taylorist Toyotist
Public-in- Transition Total
1992–
Simple control 4 (n 5 10)
4 (n 5 14)
7 (n 5 18)
0 (n 5 2)
15
Neo-Taylorist 4 (n 5 8)
21 (n 5 77)
11 (n 5 39)
7 (n 5 25)
43
Toyotist 1 (n 5 1)
6 (n 5 19)
12 (n 5 43)
4 (n 5 19)
23
Public-in-Transition 1 (n 5 3)
6 (n 5 29)
7 (n 5 26)
6 (n 5 38)
19
Total 9 37 37 17 100
Percentage of Employees
1998–
Simple Control Neo-Taylorist Toyotist
Public-in- Transition Total
1992–
Simple control 2 3 3 1 9 Neo-Taylorist 1 21 10 6 38 Toyotist 0 5 11 6 22 Public-in-Transition 1 8 9 13 31 Total 4 37 33 26 100
Notes: * Figures are weighted with two different weighting sets: that used in the upper panel allows the analyst to infer proportions in the workplaces population; that used in the lower panel allows the analyst to infer the proportions in the corresponding employee population. each of these two sets is defined as the mean of the 1992–1993 and the 1998–1999 weights. Field: Workplaces with 50 employees or more in the non-agricultural trading sector. Source: “management Questionnaire,” 1992–1993 and 1998–1999 panel sample, (n 5 371), rePoNSe surveys, Dares.
Consonant with the decline in the inten- sity of organizational change observed dur- ing the most recent period, the changes appear more significant between 1992– and 1998–1999 than between 1998 – and 2004–2005. over this latter period, the overall configuration appears to have stabi- lized and the reduction in working hours (from 39 to 35 hours per week) has rein- forced the specificity of each model rather than having altered it (Coutrot 2006). A pro- jection on the first factorial plan, represent- ing the center of gravity of workplaces present over the different survey waves, con- firms these observed trends: the spot moves from the northwest quadrant (the area where the most neo-Taylorist workplaces are projected) towards the southeast quadrant (where we find the most Toyotist-type work- places). During the two sub-periods that separate the three surveys, the direction of
change shifts slightly and its magnitude di- minishes. Thus, with the stable weight of the neo-Taylorist model, the decline in the par- ticipation of other socio-productive models (simple control and public-in-transition) im- plies to some extent a Toyotization of France’s productive base.
Renewal and Aging, contradictory trends: towards a “Re-taylorization” of Work?
In order to clearly define the dynamics at work, we used the survey’s longitudinal di- mension, based on the follow-up of a signifi- cant number of workplaces from one survey edition to the next (“perennial” workplaces). Figures 2 and 3 exhibit the net flows of work- places from one model to another (indicated by arrows) and the variations in the number of employees in each model. These two sta-
804 INDUSTRIAL AND LABOR RELATIONS REVIEW
occurred because some simple control work- places crossed critical size thresholds or they underwent significant changes in terms of both privatization and mergers and work or- ganization and industrial relations. From one period to the next, however, the dynam- ics of the two innovative models reversed. A clear Toyotization occurred between 1992–1993 and 1998–1999. According to the panel, 5% of employees moved from a neo- Taylorist structure to a Toyotist structure be- tween these two dates. Between 1998– and 2004–2005 the opposite was true; work- places shifted from the Toyotist model to the neo-Taylorist model, albeit with a modest flow of 1% of panel employees. This change nevertheless reinforces those flows related
to the decline of the simple control and, more especially, the public-in-transition mod- els. The Toyotist model was no longer pref- erable for existing workplaces; it was replaced by the neo-Taylorist model. Finally, over the two periods, the direction toward which workplaces that fit the public-in-transition model were moving is the same: the neo- Taylorist model. Departing from the public model implies a demise of work collectives and a decline in employees’ professional au- tonomy. The way this transition occurred may not only explain the acrimonious reac- tions of the employees of ex-public sector companies but would also tend to strengthen the most radical labor unions (Denis and Jeannot 2005).
Figure 2. A Toyotization Between 1992–1993 and 1998–
Notes: *All figures are weighted (by the mean of the 1992–1993 and the 1998–1999 weights) so that they correspond to the change between or within each socio-productive class (in percentage of employees in the panel): the net flow of workplaces from the simple control class to the neo-Taylorist class represents 2% of the total number of employees in the panel; given all the net flows, the neo-Taylorist class loses the equivalent of 1% of the employees of perennial workplaces between 1992–1993 and 1998–1999. For more details see Table 6. Field: Workplaces with 50 employees or more in the non-agricultural trading sector. Source: “Management Questionnaire,” 1992–1993 and 1998–1999 panel sample, (n 5 371), REPONSE surveys, Dares.
SOCIO-PRODUCTIVE MODELS IN FRANCE 805
Conclusion and Discussion It is important to keep in mind that our sample size is limited and that the indi- cators used in our analysis are synthetic; our findings thus require validation from other sources or methodologies. Nevertheless, they appear to be consistent with the results of the 2004 French survey on working condi- tions (Bué et al. 2007), and with recent field observations (Barisi 2004; Ardenti, Gorgeu, and Mathieu 2007), both of which indicate a decline in employees’ autonomy at work between 1998 and 2005. It appears as though, having passed the most intensive period of innovation, management partially retrieved
the autonomy it temporarily granted to employees. On the whole, we do not observe a gener- alized and systematic convergence towards innovative models, but rather a dual move- ment. On one hand, the renewal of the pro- ductive base definitely has contributed to the development of the Toyotist model, but only between 1992–1993 and 1998–1999; on the other hand, the aging and privatization of firms appear to have boosted the neo- Taylorist model. To some extent, the simple control model appears just as transitory— developing workplaces either fell into the Taylorist model or shifted toward the Toyotist
Figure 3. A Re-Taylorization Between 1998--1999 and 2004--
Notes: *All figures are weighted (by the mean of the 1998–1999 and the 2004–2005 weights) so that they correspond to the change between or within each socio-productive class (in percentage of employees in the panel): the net flow of workplaces from the simple control class to the neo-Taylorist class represents 1% of the total number of employees in the panel; given all the net flows, the neo-Taylorist gains the equivalent of 4% of the employees of perennial workplaces between 1998–1999 and 2004–2005. For more details see Table 7. Field: Workplaces with 50 employees or more in the non-agricultural trading sector. Source: “Management Questionnaire,” 1998–1999 and 2004–2005 panel sample, (n 5 742), REPONSE surveys, Dares.
