Article body

Management innovations are often described as playing a critical role in organisational performance and renewal (Khosravi et al., 2019; Damanpour & Aravind, 2012; Damanpour, 2020), even more so since the beginning of the Covid-19 crisis (Neeley, 2021). Management innovations are defined broadly as “the introduction of management practices, processes and structures that are intended to further organisational goals” (Volberda et al., 2013, p. 1; Birkinshaw et al., 2008). However, empirical studies on the effects of management innovations are still scarce, particularly with respect to social performance and other non-economic dimensions (known as “soft outcomes”, see Walker et al., 2015). Social performance can be understood as the extent to which employee expectations are or are not being met (as measured through indicators like employee satisfaction or retention rates, along the lines of Bocquet et al., 2019).

Liberation management, which amounts to adopting an “organisational form in which employees have complete freedom and responsibility to take actions that they, not their managers, decide are best” (Getz, 2009, p. 34), is a popular management innovation (e.g., Fox, 2020; Hamel & Breen, 2007). An ever-increasing number of companies, especially in France and in Belgium (e.g., Michelin, Kiabi, Décathlon), have been adopting liberation management (Ramboarison-Lalao & Gannouni, 2019). At the core of this innovation lies the ambition to address a social performance gap through employee empowerment (Getz, 2009; Picard, 2015). Liberated companies can thus be viewed as a typical management innovation with respect to social performance.

However, the extant empirical literature does not seem to reach a consensus on the alleged positive effects of liberation on social performance either (see Appendix B). While some claim there are clear benefits in the form of increased quality of work life (Corbett-Etchevers et al., 2019) or greater well-being at work (Ramboarison-Lalao & Gannouni, 2018), others highlight potential dark sides, such as mental exhaustion (Picard & Islam, 2019). A plausible explanation for these mixed results could be the lack of agreement on how to measure the effects of liberation management on social performance, with some studies focusing on objective criteria (e.g., turnover rates, see Hamel & Zanini, 2016) and others on more subjective criteria (e.g., satisfaction at work, see Getz, 2009).

In short, there is undeniably a need to investigate the effects of management innovations on social performance. Accordingly, our research question can be stated as follows: what can we learn from liberation management, a typical management innovation when it comes to addressing a social performance gap? In line with a growing consensus in both literatures (e.g., Bouville & Alis, 2014; Bocquet et al., 2019; Warrick et al., 2016; Battistelli et al., 2023), we operationalise liberation management by looking at its core practices (six in total). We then investigate their effects on social performance in a quasi-experiment study comparing two units (one liberated and another non-liberated) of Thermocompact, a French industrial company. This should allow us to tackle the difficult task of isolating the contingencies related to internal and/or external contextual factors. Our results are based on a survey asking questions on four variables: working conditions, loyalty, satisfaction, and happiness at work.

The paper is structured as follows. We start by presenting our conceptual framework based on the management innovation and liberation management literatures, before introducing our quasi-experimental quantitative methodology. We then go over our main findings, stressing the positive effect of three management practices on social performance: participative decision-making, personalised support, and the right to make mistakes. We conclude with a discussion of these results, of our main contributions and of three limitations sketching potential avenues for future research.

Conceptual framework

Management innovations and their contrasting effects on performance

Adopting a management innovation to address a performance gap

Studying the impact of adopting an innovation (whether managerial, technological, etc.) is by no means a new topic of interest (Camisón & Villar-López, 2014; Damanpour, 2020). Yet, there remain important theoretical questions surrounding the adoption of management innovations specifically, notably concerning their potential effects on performance (Damanpour, 2020; Walker et al., 2015). The extant literature posits that a management innovation is adopted by an organisation to harness its benefits (for its members and/or for society at large). That is, adopting a management innovation is usually viewed as the result of a strategic decision based on expected results, with several authors including targeted outcomes in their definition. For instance, Birkinshaw et al. (2008) and Volberda et al. (2013, p. 1), define management innovation as “the introduction of management practices, processes and structures that are intended to further organisational goals”.

This commonly accepted definition rests on the assumption of human rationality lying at the core of the theory of the firm, an economic school of thought which frames the adoption decision as an efficient choice resulting from a cost-benefit analysis (Volberda et al., 2014) and identification of a “performance gap” (Zaltman et al., 1973), i.e., “the perceived difference between an organisation’s potential and actual accomplishments” (Damanpour, 2020, p. 222). Management innovation is thus seen as central to maintaining and/or improving organisational performance and efficiency (Damanpour & Aravind, 2012), while ensuring sustainability in a context of uncertainty (Hollen et al., 2013). This can sometimes result in a “pro-innovation bias” (Rogers, 1995), leading organisational leaders “to take risks of unintended, unexpected, and undesired outcomes and allocate resources to innovation” (Damanpour, 2020, p.7) as they view it in a highly positive light. Empirical studies are therefore needed to clarify the outcomes management innovations can effectively bring about, whether economic or non-economic.

Hard vs. soft outcomes

There are two types of outcomes associated with the adoption of a management innovation (Walker et al., 2015). Hard outcomes refer to economic results or operational performance, for instance in terms of labour productivity or profitability. Research is primarily focused on these purely economic effects (e.g., Camisón & Villar-López, 2014). Hard outcomes further include competitive advantage and its impact on financial and business performance (Besbes et al., 2013). Another strand of research focuses on the effects of management innovations on other types of innovation, showing contrasting results: Mol and Birkinshaw (2013) find a positive impact on product innovation, while Gunday et al. (2011) identify adverse effects.

The management innovation literature is also interested in so-called soft outcomes having to do with non-economic performance. Soft outcomes are still largely understudied, as evidenced by Khosravi et al.’s (2019) meta-analysis conducted in line with Luk et al. (2008); this is confirmed by Walker et al. ’s (2015) meta-analysis showing that only 19 empirical studies out of 52 include a non-economic dimension of performance (such as customer satisfaction, formation of alliances with external players and reinforcement of legitimacy). Even fewer papers are dedicated to our topic of interest, namely, social performance (or to what extent employee expectations are being met or not) by looking at variables such as quality of work life, happiness at work or employee satisfaction (see appendix A). Some find a positive effect (e.g., Camison & Villar-López, 2010): Jiménez-Jiménez and Sanz-Valle (2011) conducted a quantitative study using objective measures to conclude that management innovations can have a positive impact on absenteeism and turnover rates; in the same vein, Wong et al. (2011) show that decentralised organisations can lead to gains in social performance. Other studies indicate a negative effect: for instance, Bouville and Alis (2014) show that lean practices may induce adverse effects on social performance, as measured subjectively. Lastly, some studies reveal no effects or only under certain conditions: for example, Bocquet, Dubouloz and Chakor (2019) draw on three case studies of French industrial firms to show that the effects of lean practices on worker health should be assessed using a combination of lean and Human Resource Management practices that differ along the lean process.

All in all, these contrasting results reveal two major concerns. First, more empirical work is needed to clarify the effects of management innovations on social performance. Second, there are some serious methodological issues surrounding construct measurement (of both management innovations and their outcomes). No consensus has yet emerged on how to measure social performance in particular (see appendix A). To fill this gap, we suggest examining more closely the adoption of liberation management, which can be viewed as a typical management innovation when it comes to improving social performance. As Černe et al. (2016) put it, focusing on “employees” non-technological innovation, not only that of managers, […] is something that is bound to become even more significant in light of the evolution of contemporary organisational structures towards more horizontal, self-managed (participatory) forms’ (p. 82) to which liberation management belongs. Besides, employees are at the core of this organisational form in which their satisfaction and happiness is clearly targeted (Getz, 2009). Accordingly, we will argue in the remainder of this paper that specific objectives need to be stated behind the adoption of a management innovation (e.g., increased social performance through employee empowerment in the case of liberation management) so that adequate measures of success can be determined (incl. indicators of employee satisfaction or happiness at work for liberation management).

Liberation management: a typical management innovation when it comes to addressing a social performance gap

Employees are widely recognised in the liberation management literature as the main recipients of this management innovation (Picard, 2015). Liberation management can thus be viewed as a means to address a social performance gap. This is openly admitted by many top managers who make the decision to adopt liberation practices: so-called “liberating leaders” essentially commit to reshaping workplace relations towards more horizontality and greater empowerment for all (Getz, 2009). This is often done in the hope to improve employee satisfaction and, ultimately, retention rates (Warrick et al., 2016). Some liberating leaders even go a step further and justify these social gains as a lever of economic performance – whether this is the case or not is another important debate, which goes beyond the scope of our paper however. Take Jean-François Zobrist, the founder and ex-CEO of Favi, a famous liberated company. He equates liberation management with “making employees happy, [which] is very easy! They need to know why they’re doing something and have the absolute freedom to do it as they see fit. […] This is the essence of the liberated company”. To which he adds: “[…] We don’t aim to make profits. Money comes from making workers happy” (Expectra, 2019, our translation).

In what follows, we will therefore consider liberation management as a typical management innovation when it comes to addressing a social performance gap, due to its commitment to empowering workers. Liberated companies have been gaining interest in the management innovation literature precisely for that reason (Fox, 2020; Hamel & Breen, 2007; Mattelin-Pierrard et Dubouloz, 2019). Accordingly, we can hypothesise that:

H1. Liberation management has a positive effect on social performance.

If top managers usually seem to have the best of intentions when adopting liberation practices, the extant empirical literature does not, however, reach a consensus on the alleged positive effects on social performance (see appendix B). Some claim liberation management is beneficial for employees notably through improving quality of work life (Corbett-Etchevers et al., 2019). Others frame it as benefits in terms of working environment, well-being, and happiness at work, induced by distinctive features of liberated companies (i.e., no pyramidal hierarchy, participatory model of decision, autonomy and mutual adjustment, see Ramboarison-Lalao & Gannouni, 2018). Some scholars have been even more critical: for instance, Picard and Islam (2019) have recently highlighted potential dark sides, manifesting as mental exhaustion, scapegoating of contrarian individuals, or experiencing normative pressures to be overly happy.

In short, as with other management innovations, there is undeniably a need to clarify the effects (and most appropriate measurement tools) of liberation management on social performance.

Studying social performance through the adoption of management practices

Now that we have established the need to further investigate the effects of management innovations (incl. liberation management) on social performance, we need to think of a way to do so. There is a growing consensus in the management innovation literature about the usefulness of looking at the adoption of management practices (e.g., Bocquet et al., 2019; Walker et al., 2015). Examining management practices appears indicated for three main reasons. First, it does justice to the complexity of management innovations by allowing us to describe them more comprehensively (i.e., as the combination of structural effects, management tools and practices). Second, management practices are well suited to compare the effects induced by different management innovations. Third, it allows us to look at tangible results, beyond mere statements of intent and promises made by leadership.

Going back to our topic of interest, this means that we will need to better delineate so-called liberation practices. Management practices are not foreign to the liberation management literature: liberation management has emerged as a concept in a context where management practices were being renewed (Gilbert et al., 2018); moreover, liberated companies have been described before as a distinct set of practices (or even bundles of practices, see Battistelli et al., 2023) aimed at improving social performance (e.g., the absence of external signs of power or hierarchical control, see Picard, 2015; Warrick et al., 2016). Drawing on the literature review by Mattelin-Pierrard (2019), we have identified 11 management practices that are commonly adopted in liberated companies (see table 1).

Table 1

Review of liberated company management practices

Review of liberated company management practices

-> See the list of tables

Some of these practices are recognised to have a positive effect, directly or indirectly, on well-being at work. Having the right to make mistakes, especially for employees facing heavy workloads, is said to increase perceived organisational support and improve self-esteem at work (Wang et al., 2020). Self-managed teams also seem to translate into social performance gains by having a positive impact on the variety of tasks performed by employees and their autonomy (Van Mierlo et al., 2005; Cohen & Ledford, 1994). Finally, a recent study indicates a positive relationship between employee participation in its different forms (i.e., participation in management, ownership, and company results) and well-being at work (Uribetxebarria et al., 2021).

In another contribution (Battistelli et al 2023), we have argued that liberation management practices can be further divided into two categories: “action-oriented” practices and “enabling” practices. The former includes practices such as self-direction or self-management, more directly aligned with the core ambition of liberation management (i.e., empowering employees to improve social performance). Enabling practices, such as information transparency or tolerance for mistakes, can be said to contribute to this goal more indirectly, as they are meant to help employees act by supporting their initiative. Based on this distinction, we further hypothesise that:

H2. “Action-oriented” practices (e.g., self-direction or self-management, see Battistelli et al., 2023) have a stronger positive impact on social performance than “enabling” practices.

Methodology and data

Our objective is to examine to what extent liberation management may affect social performance. To tackle this question, we came up with a two-step methodology: 1) a combination of a literature review and expert interviews to delineate the core practices of liberated companies (six in total); 2) a quasi-experiment to study the effects of these core practices on social performance.

Expert interviews

In July and August 2018, we conducted 9 semi-structured expert interviews (for a total of 10 hours 40 minutes and 159 pages of transcripts). These were conducted following the qualitative Delphi method (e.g., Sekayi & Kennedy, 2017). Because liberated companies are widely publicised, this method, which gives a voice to a wide range of actors and “encourages honest opinion [,] free from peer group pressure” (Williams & Webb, 1994, p.181), appears to be particularly indicated in the case of liberation management. We also thought it would be suitable for checking the consistency between the practices discussed by the experts interviewed with the practices identified in the literature (see table 1).

The interviews were conducted in person, over the phone or online. We selected the interviewees based on two criteria: first and foremost, their acquaintance with liberation management; second, their different perspectives on the subject (representatives of the industry vs. academia). Our two-part interview guide included open-ended questions (verifying expert status through their acquaintance with liberated companies), and scale questions integrating 35 statements on management practices. Four mutually exclusive answers were possible for each statement: “this practice is: (1) Antinomic with liberated companies (this practice cannot be present in a liberated company)/ (2) Not representative of liberated companies (this practice may be present in a liberated company but is not distinctive)/ (3) Representative of liberated companies (this practice is present and distinctive)/ (4) Inseparable from liberated companies (without this practice liberated company status cannot be granted)”. To be considered a core liberation practice, at least eight experts needed to view it as “inseparable” from liberation management.

Ultimately, by combining Mattelin-Pierrard (2019) literature review and these expert interviews, we were able to retain six management practices as “core liberation practices” (out of the 11 listed in table 1). Three fit into the category of “enabling” practices (i.e., the right to make mistakes, personalised support and information transparency) introduced above, while the remaining three can be viewed as “action-oriented” practices (i.e., self-direction, self-managed teams and participative decision-making).

Quasi-experiment

It appears from our literature review that there is no quantitative study on the impact of liberation management on social performance. This is somewhat surprising considering the number of empirical studies dedicated to the effects of management innovations more generally. To fill this gap, we decided to conduct a quasi-experiment in a partially liberated company, to “[back up] causal inference while maintaining internal and external validity without interrupting ‘real life’ through intrusive intervention” (Grant & Wall, 2009, p. 655). Quasi-experiments are well suited for performance evaluation as they come with no requirements in terms of performance levels and are able to control for contextual factors, which are known to have a significant impact on performance. In the liberation management literature more specifically, it is often stressed that benefits resulting from the adoption of liberation practices might have to do with internal or external factors such as the company’s size, culture or economic sector (Ramboarison-Lalao & Gannouni, 2018; Corbett-Etchevers et al., 2019) or the centrality of the “liberating leader” (Carney & Getz, 2009; Chabanet et al., 2017; Gilbert et al., 2017a).

We chose to study Thermocompact, a French industrial firm that is part of the Thermo Technologies group (with a turnover of €89 million in 2017 and a total of 330 employees). Our quasi-experiment took place in one of its factories located in Haute-Savoie which is divided into two business units (BU). The wire electrical discharge manufacturing (EDM) unit (described in what follows as the liberated unit) started implementing liberation management practices in 2013; the surface coating (SC) unit (described hereafter as the non-liberated unit) preferred not to adopt liberation management. In other words, leadership left the BUs free to take up liberation practices or not. Initially, discussions were held with the EDM unit with the goal of co-constructing the “company’s future”. Liberation management quickly became an object of interest, with several discussion sessions organised around its practices. The employees in the SC unit decided not to get on board as they were under the impression that they would need to adopt the practices discussed by the EDM unit necessarily. It seems they did not understand the intention to “co-craft” organisational changes and so, for the sake of consistency, leadership preferred not to “impose” the adoption of liberation practices onto them. Being a partially liberated plant, Thermocompact thus constituted a great case study for our quasi-experiment: by comparing its two BUs, we were able to isolate the effects of liberation practices on social performance (i.e., effects that can be directly imputed to typical liberation practices, see Ramboarison-Lalao & Gannouni, 2018).

The first part of our quasi-experiment consisted of: a semi-structured interview with the CEO (1 hour 29 minutes), informal discussions with the CEO and the human resources director, a tour of the plant and non-participant observation of eight meetings (January 2018 to September 2019). We then asked the employees from both BUs to fill out a questionnaire structured as follows: first, they were asked to indicate how frequently they apply the six core liberation practices identified above; second, they were asked to rate the social performance of their unit on four measures (working conditions, loyalty, happiness and satisfaction at work); third, they were asked about two contextual variables, namely, their unit affiliation (in order to check the coherence between labels and actual practices) and perception of client pressure (as the preliminary interview with the CEO showed that the non-liberated unit was subject to more client pressure mainly because of shorter delivery times).

The questionnaire was administered to workers on site in November 2018. The factory employs 143 people and operates three production shifts (i.e., early morning, evening, and night shifts). In total, 109 workers completed our survey: 50 in the non-liberated unit and 59 in the liberated one, giving us a 76.22% response rate. The respondents came to a conference room in groups of three or four to complete the questionnaire. We went for a self-self-administered questionnaire as these usually attract more honest responses, thereby reducing social desirability bias (Alt et al., 2015). To preserve the anonymity of respondents, only aggregate results were presented to Thermocompact’s CEO and human resources director as well as to the human resources manager of the factory studied.

Description of variables[1]

Dependent variable – social performance – The dependent variable social performance is the mean score of four variables, namely: working environment, happiness at work, satisfaction at work and employee loyalty. As already discussed, it is widely admitted in the literature that liberation management is intended to address a social performance gap (i.e., to increase social performance through employee empowerment). However, it is still unclear to what extent it is successful in doing so. A plausible explanation for the mixed results presented above could be the lack of agreement on how to measure the effects of liberation management on social performance. To our knowledge, the evaluation tools brought forward in the literature have not yet been reviewed systematically. Appendix B summarises our attempt to fill this gap, showing there exists a diverse range of indicators, with some studies focusing on objective criteria (e.g., turnover rates, see Hamel & Zanini, 2016) and others on more subjective criteria (e.g., satisfaction of personal needs, see Getz, 2009).

Despite this lack of consensus on what exact criteria to use, it appears that all the indicators proposed so far are to some extent concerned with employee empowerment. This suggests that social performance may be best assessed in a tailored manner, that is, by using indicators that may not be as relevant for other management innovations (e.g., lean management). For liberation management, we have identified four key variables (again, working environment, happiness at work, satisfaction at work and employee loyalty) after reviewing existing studies on the social performance of liberated companies (Corbett-Etchevers et al., 2019; Ramboarison-Lalao & Gannouni, 2018) and the commitments traditionally made by “liberating leaders” (in line with the “performance gap” perspective on management innovations introduced in the previous section).

To collect data at the organisational level, we integrated these variables into an employee survey, which is recognised to be an effective and efficient way to gather information on perceptions and satisfaction levels (Rogelberg & Stanton, 2007). Employees were asked to rate on a four-point Likert scale (reversed for analysis) their responses: “Concerning the following statements[2], in your unit do you 1= Totally agree, 2= Agree, 3= Disagree and 4= Totally disagree?” [3] Subjective measures like employee perception or satisfaction are widely used in innovation-performance studies (Damanpour, 2020; Walker et al., 2015): “most of the data collected in innovation surveys are qualitative, subjective and censored” (Mairesse & Mohnen, 2010, p.8) and “many variables, whether qualitative or quantitative, are subjective in nature, based largely on the personal appreciation and judgement of the respondents.” (Pinget et al., 2015, p.133). This is even more true for the evaluation of social performance and other soft outcomes, as summarised in appendix A.

The survey was administered to employees for two main reasons. First, we have already seen how they can be regarded as the main recipients of liberation management (Picard, 2015). At the risk of repeating ourselves, liberation management is aimed at filling a social performance gap through employee empowerment. Second, the literature seems to overlook that clear ambition by hardly ever giving a voice to employees and focusing largely on the opinion of liberating leaders instead.

Lastly, the internal consistency of our dependent variable (social performance) was checked and validated using Cronbach’s . Our value was 0.770, which is considered acceptable by most standards (> 0.7, see Nunnally & Bernstein, 1993).

Independent variables – six management practices – The degree of implementation of the six core liberation practices was measured using 15 variables derived from existing empirical studies. Mean scores were calculated for all the variables included in each practice (as detailed in appendix C). The variables were introduced by the following statement: “Evaluate the frequency of implementation of the following practices in your unit” and rated on a three-point scale (reversed for analysis): 1 = Always, 2 = Sometimes and 3 = Never. We made sure to have a reasonable scale width (i.e., the number of response options) to decrease the non-response risk (Rogelberg & Stanton, 2007), knowing that this would not affect Cronbach’s in a single study, especially with three response options (Matell & Jacoby, 1972; Voss et al., 2000).

Control variables – Again, quasi-experiments present the advantage of controlling for a certain number of contextual factors. In our case, these were reduced to perceived client pressure (mean score of), as rated on a four-point Likert scale, and a dummy variable indicating whether the respondent belongs to the liberated or non-liberated unit.

Results

Descriptive statistics

The liberated (50 employees) and non-liberated unit (59 employees) distribution was fairly equal. We conducted a preliminary analysis to check for violations of assumptions of normality (skewness, kurtosis and normality plot), linearity (scatterplots), multicollinearity (collinearity diagnostic, variance inflation factor < 4 and tolerance > 0.2) and homoscedasticity (scatterplots).

Table 2 presents the descriptive statistics of our variables. The relationship between liberation status, the six core liberation practices and social performance was investigated using Pearson correlations (see table 3). Pearson coefficients appear below the diagonal and the results of our test for statistical significance above the diagonal. Examination of the correlations reveals that affiliation with the liberated unit is significantly and positively correlated with social performance with a p-value<0.001 (r=0.424, n=109). The results further indicate that client pressure is significantly but negatively correlated with social performance (r=-0.339, n=106, p<0.001). As expected, all six core liberation practices are positively and significantly correlated with social performance, i.e., the right to make mistakes (r=0.374, n=108, p<0.001), self-direction (r=0.349, n=109, p<0.001), self-managed teams (r=0.533, n=109, p<0.001), participative decision-making (r=0.456, n=108, p<0.001), personalised support (r=0.559, n=107, p<0.001) and information transparency (r=0.479, n=109, p<0.001).

Table 2

Descriptive statistics of the variables

Descriptive statistics of the variables

-> See the list of tables

We used an independent samples t-test to identify potential differences in social performance and implementation of management practices between the liberated and non-liberated units.[4] The fourth column of table 4 presents our t-test values, showing that: there is a significant difference between the two units with respect to social performance (p<0.001) and client pressure (p<0.001); the implementation rate of the management practices studied was higher in the liberated unit, as we expected; except for the right to make mistakes, for which no significant differences were observed between the two units.

Table 3

Pearson correlations and significance

Pearson correlations and significance

** Significant at .01 level. * Significant at .05 level.

-> See the list of tables

Table 4

t-test between liberated and non-liberated units

t-test between liberated and non-liberated units

** Significant at .01 level. * Significant at .05 level.

-> See the list of tables

Hierarchical multiple regression

We ran a hierarchical multiple regression to assess the ability of the six core liberation practices to predict social performance, while controlling for the effects of perceived client pressure and liberation status. Using a p<0.001 criterion for standardised residuals, we found no outliers i.e., no observations exceeding, in line with Tabachnick and Fidell (2012).

Two models were evaluated, with social performance as the dependent variable in both cases. Model 1 includes two control variables: perceived client pressure and, in line with our quasi-experimental design, the dummy variable liberated (i.e., whether the respondent belongs to the liberated unit or not). Model 2 further includes the six core liberation practices. Appendix D summarises the variables used in our hierarchical regression. Table 5 displays the standardised coefficient as well as R2 and adjusted R2 with a 95% confidence interval. The regression coefficient is significantly different from zero, i.e., F(2, 101)=15.412, p<0.001 with R2 at 0.219 for model 1 and F(8, 95)=11.438, p<0.001 with R2 at 0.518 for model 2, thereby indicating that our models fit the data well. The adjusted R2 value of 0.518 suggests that model 2 (incl. management practices and our two control variables) can predict more than half of the value of social performance, with a p-value<0.000. The corresponding F variation also appears to be significant (at p<0.000). Therefore, it seems that the inclusion of the six core liberation practices in the model helps to explain 29.9% of social performance scores, even after accounting for perceived client pressure and liberation status.[5]

Table 5

Hierarchical regression analysis on social performance[6]

Hierarchical regression analysis on social performance6

Standard error in brackets.

*** Significant at the .01 level. ** Significant at the .05 level. * Significant at the .1 level.

-> See the list of tables

The results of both models indicate a significant and positive effect (p<0.05) of liberation status on social performance. This is consistent with the t-test results, which, again, showed significant differences in social performance between the two units. Perceived client pressure also has a significant, but negative, impact on social performance (p<0.01).

Model 2 allowed us to test our hypotheses (i.e., whether liberation practices can have a significantly positive impact on social performance (H1) and some of them – “action-oriented” practices – more so than others – “enabling” practices (H2)). Our results indicate that three out of the six management practices identified as core liberation practices are predicted to have positive effects on social performance: the right to make mistakes (p<0.001), participative decision-making (p<0.05) and personalised support (p<0.001). The significance of these results was confirmed using 95% confidence intervals, which, except for self-managed work teams (p<0.1), did not include zero as a possible value. By contrast, we saw no evidence of significant positive effects for the practices of self-direction and information transparency.

In summary, H1 is supported by our study: we have witnessed an increase in social performance following the adoption of liberation management in the corresponding unit. H2, however, is only partially supported by our results: we would have expected all three “action-oriented” practices (i.e, participative decision-making, self-direction and self-management) to have a positive impact on social performance but found that only one of them does (i.e., participative decision-making).

As for the practices for which we found no significant effects (i.e., information transparency, self-managing teams and self-direction), perhaps we should note that, in another paper (Battistelli et al. 2023), we have argued that the implementation of liberation practices tends to be more sustainable when these are adopted in bundles displaying internal and temporal coherence. Accordingly, nothing excludes that we could have found an interaction effect of some of these practices, had we been able to test this hypothesis with a bigger sample size.

Discussion and conclusion

This research aimed to cast light on the effects of management innovations on social performance, which tend to be too often overlooked in the literature. To do so, we chose to examine more closely the adoption of liberation management because employees are widely recognised to be the main recipients of this innovation (Picard, 2015). Therefore, liberation management can be viewed as a typical management innovation in terms of improving social performance. We drew two hypotheses from the liberation management literature: H1 holds that there should be a positive relationship between liberation management and social performance, while H2 contends that, amongst the practices that can be regarded as core liberation practices, so-called “action-oriented” practices should have a stronger impact than so-called “enabling” practices (Battistelli et al. 2023). Our results indicate that three out of the six management practices identified as core liberation practices have a clear positive effect on social performance, namely: participative decision-making (“action-oriented” practice), personalised support (“enabling” practice), and the right to make mistakes (“enabling” practice), thereby supporting H1, and H2 only partially. Before we discuss these results in greater detail, let us make some preliminary remarks.

Some preliminary remarks on contextual factors

As Grant and Wall (2009) recall, the context in which a quasi-experiment takes place may have an influence on some of the observations made. In our study, we believe that contextual factors may be responsible for two findings: the fact that a core liberation practice, the right to make mistakes, is implemented in both units, and the common perception of significant client pressure with adverse effects on social performance (i.e., only partially mitigated by the adoption of liberation management).

Concerning the former, we can assume that the right to make mistakes has unintentionally diffused to the non-liberated unit because both units are led by the same CEO. This seems plausible, given that the right to make mistakes is usually one of the first practices to be adopted in the liberation process (shortly followed by other so-called “enabling” practices, e.g., information transparency, personalised support, see Battistelli et al. (2023). As for the experience of client pressure with ill effects on social performance in both units, two things can be said. First, the units occupy different positions in the value chain. While the liberated unit acts as a supplier, the employees in the non-liberated unit work as subcontractors, with arguably more pressure on their shoulders (i.e., due to the potential ripple effects of their actions on the rest of the value chain). Moreover, the liberated unit may feel better equipped to deal with the whims of their customers, given that liberation practices actively contribute to their empowerment. The practices of self-management and personalised support seem particularly well adapted to relieve them of extra pressures potentially exerted by managers in the other unit.

The remainder of our discussion will be centred on three takeaways.

The relevance of subjective criteria for soft outcomes

A first point of discussion concerns our proposed methodology, stressing the relevance of subjective criteria for measuring soft outcomes (see Khosravi et al., 2019), and more specifically, the social performance of management innovations. This is particularly true in the case of liberation management considering that, again, employees are recognised to be the main recipients of this approach (Picard, 2015). It makes sense, then, to take their view into account when assessing the success of that management innovation. Now, one could object that relying primarily on subjective data does not counter the risk known as “adaptive preferences” in a satisfactory manner (see e.g., Mitchell, 2018). Arguably, there may be instances where people’s preferences are not really their own, due to them having internalised some external, social pressure[7]. This is a serious objection, especially bearing in mind the potential dark sides of adopting liberation practices (such as the experience of normative pressures to be overly happy, see Picard & Islam, 2019). If such a case can be made for supplementing subjective criteria with more objective measures (e.g., turnover rates), by no means does this call into question the relevance of our methodology. Subjective data would still be needed to make sense of more objective measures because the whole point is to investigate to what extent employee expectations are or are not being met. This is even truer for liberation management. In line with their core ambition to empower employees, we may wish to avoid another risk, that of paternalism, and consider people to be the best judges of their own welfare. Although it may seem somewhat counterintuitive at first, higher turnover rates observed in some liberated companies (see Hamel & Zanini, 2016) could go hand in hand with greater satisfaction levels. The best way to know why there tends to be such an increase is to ask employees to share their reasons for leaving. It may well be that some of them are not resigning because they are burnt out, but because they are now empowered to the extent that they feel ready to start their own business.

No “one best way” to measure the social performance of a management innovation

Secondly, we would like to stress how our mixed results reinforce the importance of being clear about the objectives and measures of success before adopting a management innovation aimed at improving social performance (such as liberation management). To make sure the adoption is both sustainable and efficient, social performance may be best approached in an idiosyncratic manner: that is, by focusing on the “performance gap” (Zaltman et al., 1973) or specific “organisational goals” (Birkinshaw et al., 2008, p. 825) addressed by the innovation management in question (e.g., greater happiness at work through employee empowerment in the case of liberation practices). In other words, before adopting a specific management innovation, the main ambitions behind its implementation should be stated as clearly as possible to establish adequate measures of success. This means that there is no single “one best way” to assess the social performance of a management innovation. This is no easy task as several goals may be pursued at once, top managers may not be fully aware of their motivations and/or they may deem some of these ambitions socially undesirable. For instance, others (Gilbert et al., 2018) have shown that top managers tend to adopt liberation management also for legitimacy purposes (towards internal and/or external stakeholders).

These findings contribute to a better definition of the concept of management innovations, by reinforcing the importance of intentionality in their adoption (Dubouloz, 2013). Intentionality refers to the ambition to increase performance (Le Roy et al., 2013; Mol & Birkinshaw, 2009) while achieving organisational goals (Birkinshaw et al., 2008; Volberda et al., 2013). In other words, it may be argued that management innovations do not encompass all the management practices newly implemented in an organisation but only the ones adopted with clear objectives in mind – whether realised or not (Damanpour & Aravind, 2012). This is a key, yet often neglected, aspect, with only 11 studies out of the 25 most popular ones integrating intentionality in their definition of management innovations (Mattelin-Pierrard 2019). Even those papers that explicitly address this dimension do not seem to reach a consensus on the nature or scope of intended goals: examples given are “creat[ing] value for the organisation” (Damanpour & Aravind, 2012, p. 424), “improv[ing] the effectiveness and efficiency of internal organisational processes” (Dubouloz, 2013, p. 47) or “produc[ing] changes in the organisation’s managerial procedures and administrative systems” (Damanpour, 2014, p. 1269). Our results allow us to clarify two points: (1) the adoption of a management innovation may generally be regarded as “intended to further organisational goals” (Birkinshaw et al., 2008, p. 825), whatever these may be; (2) these goals tend to vary from one management innovation to another, thereby justifying the need to specify the objectives behind their adoption (e.g., increased social performance through employee empowerment in the case of liberation management) and define tailored measures of success (incl. indicators of employee satisfaction or happiness at work for liberation management).

A perception gap behind the social performance gap: a pro-liberation bias reinforced by a strong rhetoric

Thirdly and finally, this difficulty to clarify the organisational goals behind the adoption of liberation practices may help explain the differing effects between our six practices of interest. The frequent confusion surrounding the aforementioned social performance gap may thus lead to different interpretations by top managers vs. employees (i.e., a perception gap). As mentioned hereinabove, top managers may seek to strengthen their legitimacy, thereby feeling the need to justify the adoption of liberation practices with a strong rhetoric: they are proposing a new narrative about work relations and may be tempted to boast about their good intentions to make it more convincing. This resonates with Gilbert et al.’s (2017a) depiction of liberation management as yet another trend that can serve as an inspiration for company leaders to initiate change. In a similar vein, Casalegno (2017, p. 240, our translation) speaks of this management innovation as “a rhetorical device conveying the necessity to change without clearly stating the nature of that change”. Top managers may thus be primarily concerned with inspiring change, or setting employees in motion, when adopting liberation practices. This may come at the cost of overcommitment, thereby leading to feelings of frustration or, worse, disenchantment amongst employees (see e.g., Fox & Pichault, 2017; Battistelli et al., 2023). It may well be, then, that employees at Thermocompact pay less attention to the three practices of information transparency, self-managed teams and self-direction because they find the efforts behind their adoption wanting. For instance, they may have high expectations in terms of salary transparency, which is rarely put into place in liberated companies (with the notable exception of Semco).

A “pro-innovation bias” may thus be visible not only amongst top managers, as highlighted in the management innovation literature (Damanpour & Aravind, 2012; Rogers, 1995), but also amongst employees. Put differently, the strong rhetoric cultivated by top managers may lead employees to formulate higher expectations towards liberation management. This may be particularly true with so-called “action-oriented” practices (such as self-direction and self-management according to our results), more directly aimed at employee empowerment. The implementation of those practices may be more easily taken for granted, thereby not causing as much satisfaction amongst employees. Participative decision-making seems to form an exception, as this “action-oriented” practice is associated with a positive effect on social performance in our study. The rationale behind this may have to do with the fact that Thermocompact’s employees have taken part in some major strategic decisions (e.g., by having a say in whether to adopt liberation practices or not), which is usually not advertised (thus not to be expected) in liberation management (Getz, 2009; Battistelli et al., 2023). Conversely, employees may be pleasantly surprised to witness so-called “enabling” practices being put into use, that is, to see their managers actively support their empowerment (by showing more tolerance for mistakes and acting as servant leaders at their disposal, according to our results). As is well known in social psychology (e.g., Van Dijk et al., 2002), lowering one’s expectations in the face of uncertain outcomes could lead to greater levels of satisfaction (or, at least, lower levels of dissatisfaction).

All in all, we are not saying that a strong rhetoric may not be useful to sustain the use of a management innovation, notably by strengthening its legitimacy (i.e., by giving meaning to its adoption, see Arnaud et al., 2016). However, top managers should be equally wary of the risk of overcommitment. Otherwise, they may negatively affect social performance by causing feelings of frustration or disenchantment amongst employees. In particular, managers should resist the urge to focus solely, or primarily even, on “action-oriented” practices simply because these are more directly aligned with the core ambition to empower workers. Instead, we would argue that managers should first devote efforts and resources to implementing “enabling” practices. Not only would this seem to render the adoption of liberation management more sustainable (Battistelli et al. 2023) but also, it could lead to substantial gains in social performance according to our results. Thus, our study reaffirms the role played by managers in successfully adopting liberation management (thanks to “enabling” practices like personalised support), while contributing to better defining the contours of that management innovation (as the necessary combination of “action-oriented” and “enabling” practices).

Contributions

Our theoretical contributions to both the management innovation and liberation management literatures are twofold. First, we stress the relevance of subjective criteria for measuring soft outcomes such as social performance. Second, through the typical example of liberation management, we argue that social performance may be best approached in an idiosyncratic manner, using tailored indicators (i.e., directly aligned with the performance gap or core ambition behind the management innovation in question, namely, happiness at work through employee empowerment for liberation management).

This research also makes a couple of empirical contributions. First, its originality lies in the quantitative method used. Quasi-experiments are rare in the field of strategic management (Walker et al., 2015) despite their clear methodological benefits (notably, with respect to isolating contingencies, Grant & Wall, 2009). This is even more true for management innovations, knowing how hard they are to operationalise (Damanpour & Aravind, 2012). Second, we show which liberation practices can actually make a difference when it comes to improving social performance, a growing concern for managers (Chabanet et al., 2017).

Finally, in terms of managerial contributions, we further show that top managers should be wary of the risk of overcommitment when communicating on a newly adopted management innovation, as highlighted before (e.g., Fox & Pichault, 2017). In the specific case of liberation practices, we have seen that a strong rhetoric may induce a pro-innovation bias amongst employees, likely to heighten expectations (particularly towards “action-oriented” practices) and cause feelings of frustration or disenchantment in the long run. However, our results suggest that two “enabling” practices (the right to make mistakes and personalised support) can mitigate this effect by playing an important part in supporting employee empowerment.

Limitations and avenues for future research

Three limitations in our study can be turned into avenues for future research. Firstly, the results should be interpreted with caution since they are based on a single case study. That is, our study should be replicated with larger samples, in other contexts and/or using longitudinal data (as perceptions may evolve over time). Indeed, we may want to try and generalise our findings to other contexts (Shadish et al., 2001) and management innovations. This should also allow us to test the possibility of interaction effects between the practices investigated (notably those for which we did not find any effects in isolation).

Secondly, future research could integrate a variety of internal and/or external factors to investigate to what extent liberation management may strengthen a company’s legitimacy. Organisational performance is often regarded as being affected by stakeholders’ perceptions of legitimacy, thereby leading organisations to seek both internal and external legitimacy notably by building a good reputation or goodwill (Orliztky et al., 2003). Liberated companies are no exceptions to this rule, with famous examples of vision statements clearly aimed at securing stakeholder support: e.g., “Sustain the lives of two hundred families in Hallencourt.” (FAVI) or “Committed to creating a better world together!” (CHRONO Flex).

Thirdly and finally, we focused here on one type of outcomes, social performance, when the liberation management literature usually understands performance multidimensionally (economic, social, environmental and societal outcomes) with potential mediating effects between these dimensions: for instance, Getz (2009) theorises a virtuous circle between employee satisfaction and “world-class” economic performance; similarly, Corbett-Etchevers et al. (2019) observe mutual reinforcement between social performance and product innovation. Future work could thus examine the mediating effects of social performance on product innovation and economic performance, when adopting a management innovation. We may well find that organisations have good reasons to “do well by doing good” rather than “do good and do well” (Wood & Jones, 1995, p. 235).