French companies’ quality assurance and performance in a crisis context

Abstract

This study examines the causality effects between quality control assurance certification and corporate performance. The study uses the matching method that Donald B. Rubin first introduced in 1974. By matching the 2006 COI survey and the DIANE database, the available data show that: while quality control assurance certification has given a major boost to French companies’ financial capital productivity, it has no effect on their added value, turnover and the labor productivity.

Keywords

Certification corporate performance selection bias matching estimator matching method

Kais Mtar, A Doctor of Economics at the University of Nice Sophia Antipolis, MTAR Kais, aged 36 got his Master’s degree in 2007 from the same institution. In 2015, he worked as Assistant Professor on contract basis at the “University of Rennes 2” in France. Apart from taking part in seminars and conferences, Kais has already published an article in the year 2015 which is entitled ‘FDI’s Effects on the Tunisian Economy in a New Context’, (2015), International Journal of Applied Research and Studies 2015, Volume 4, Issue 1, P.1-12.

Souad Bannour (born in 1983 in Mahdia, Tunisia) holds a Phd in Economic Sciences (2013) from the University of Sousse and the University of Nice Sophia Antipolis, France. She obtained her Master’s degree in 2007 from the same institution. From 2014 till now, Souad has worked as Assistant Professor on contract basis at the High Institute of Management in Sousse as well as in the Faculty of economic Sciences and Management, Mahdia. She also worked as temporary teacher at the high Institute of Technological Studies in Mahdia. Apart from taking an active part in related seminars and conferences, Souad has already published 3 papers in the year 2015: The first article is entitled: The Determinants of FDI in Tunisia: An Empirical Study through a Gravity Model, Asian Economic and Financial Review 2015, 5(12): 1306-1314. The second one is about: The Attracting Factors of Foreign Direct Investment: Tunisia’s Cause After the Revolution, Journal Advanced Research 2015, Volume 3, Issue 2, P.369-375. The third one is about: FDI’s Effects on the Tunisian Economy in a New Context, (2015), International Journal of Applied Research and Studies 2015, Volume 4, Issue 1, P.1-12.

1 Introduction

Since the late 1980’s, the theme of the effects of setting up quality insurance in companies has been the subject matter of an increasing number of authors in the academic circles. Stringent consumer demands, market globalization, fierce competition and the dissemination of ICTs (Information Communication Technology Services) revealed an ever-growing demand of companies in favor of greater integration of the quality assurance approach to manage their processes and their activities, so that their products and services are consistent with the goals they have set. As such, the implementation of quality assurance certificates by companies can be seen as an effective tool to maintain a competitive edge [33]. It is part of a dynamic of ongoing progress essential to the development of organizations and business modernization [8]. It is in this context that many French companies have engaged in various initiatives to formalize their quality assurance programs.

If these standards meet certain expectations that lead to a national and international consensus on the operational, financial and organizational benefits provided to businesses [34], they also raise concerns among a number of managers as well as fuel fears of scientific nature as to how to approach and deal with this phenomenon. One of the main issues that literature, especially sociology of work and employment [31, 32], has to deal with is the cost of certification and cumbersome procedures which may conflict with the new management of the company in terms of flexibility, responsiveness and autonomy. From an empirical perspective, this result was confirmed in the study by [29] who show that the quality assurance certification, including ISO (International Organization for Standardization) 9000, has no effect on the volume of sales and has a negative impact on financial performance. Thus the question of the real profits of quality assurance standards still continues to raise controversy, both in research and in companies despite the large number of studies on the subject [5].

In this context, the present paper attempts to provide answers to the impact evaluation of implementing quality standards on corporate performance. To solve this problem, a matching method, as originally developed by [36], has been used, which indeed offers an interesting alternative to the analysis implementing nonparametric estimates while maintaining the individual character of the effect of certification on business performance. The peculiarity of this method is its ability to estimate the overall effect of the certification only to certified companies and that this effect is homogeneous within companies, in contrast to estimates by the method of nonlinear least squares of a Translog or Cobb-Douglas production function, which requires an estimate of the global effect of certification on all businesses.

This paper is organized into five parts. While the first section provides some historical context that led to the emergence of the quality assurance program and its evolution in the corporate organization. A second part will be devoted to a literature review on academic studies that have sought to understand the link between the implementation of certification by businesses and the expected profits. The presentation of the econometric model and the estimators are outlined in the third part. The fourth part presents the data from matching the 2006 COI (Organisational Changes and Computerisation) survey and DIANE (France’s Disc for Economic Analysis) database. Finally, Section V discusses and comments on them.

2 The quality assurance program: Historical overview and presentation

The history of quality assurance program shows that the first manifestations of this mechanism date back to the Egyptians. The latter had a particularly specific quality assurance program related to the preparation of the graves. During the following centuries, this program has been the subject matter of evolutionary trends [10]. Not until the 1950s with World War II that this program is actually used in the United States in the aerospace, military and nuclear industries. However, this program has been applied to other industries in the years 1960–1970. In France, it was first introduced in 1965, in developing satellites. The transfer of the quality assurance program to other sectors (automotive, electronics and Information technology sectors in particular) began in the early 1980s, especially in companies like Renault and the PSA Peugeot Citroën (manufacturer of Peugeot and Citroën vehicle) [18].

Indeed, the quality assurance program remains a voluntary process that involves providing a guarantee to customers about the nature of the service provided through the compliance of its internal activities to a set of rigorous procedures of operation. On the one hand, this program is mainly concerned with the successful completion of each step towards the product realization, including design, production, and placing on the market. On the other hand, it deals with the relations between these different steps, particularly through the procedures of implementation and transmission of technical specifications between the design office and the manufacturing processes. Its implementation aims primarily at preventing and detecting defects and shortcomings as well as correcting them in the easiest and most cost-effective way possible. [18] defines quality assurance as “all the pre-established activities implemented to provide the appropriate confidence that a product or service will satisfy the demands of quality”

The main objective of this program is to ensure that the company making the product or service is sufficiently well organized so that the quality of its products is invariably guaranteed. The degree of structuring needed to ensure that quality is that of an organizational model called quality standard. Its purpose is to ask an organization to prove that it has implemented a quality program [10]. The corollaries of quality assurance are the certifications issued by ISO [10]. Published in 1987, revised in 1994 and 2000, the ISO 9000 standards have become essential in the management and organization of businesses. They affect all countries and all sectors of activity. These standards reflect a consensus about the essential characteristics required of a quality program to ensure the effective functioning of any organization. These organizational tools are widely disseminated since their creation in the world, especially in France. Along with these international standards, specific standards have been established to ensure better quality, the VDA 6 (Verband der Automobilindustrie) for Germany, and EAQF (Evaluation Ability Quality Supplier) for France [18], the QS (Quality Systems Requirements) 9000 for the USA (United States of America) and the EAMS (Entreprise Asset Management System) for Europe.

The quality assurance program is becoming increasingly important to meet customer requirements inasmuch as they serve as relevant tools of control. These mechanisms have long been reserved for sectors such as aeronautics, nuclear and armaments industry. The promulgation of ISO 9000 in 1987 has opened these approaches to new industries and new activity sectors. For France, the results of the COI surveys (1997 and 2006) show that quality certifications are highly disseminated within French companies. Table 1 shows that these standards have evolved significantly: the share of companies that have adopted quality certifications increased from 46.7% to 60% between 1997 and 2006. This development is spread across industries. Thus, they are highly disseminated in the coking industry, refining, nuclear industries (94.7%), and the manufacturing industry of electrical and electronic equipment (82.4%) and manufacturing industry in transport equipment (79.3%). This trend is reversed in the case of companies belonging to the textile and clothing industry (26%) as well as the leather industry and footwear (23%).

3 A literary review on the advantages of quality certification

Since its inception, most managerial studies show that the development of quality certificate yields significant benefits to companies. Literature offers an increasing number of variables weighing the contribution of the implementation of this new managerial practice. For simple analysis purposes, [33]) bring together these variables in three dimensions: the operational aspect, the financial aspect and the organizational dimension.

3.1 The operational dimension

The authors present the contributions of the certification at the operational level of the company distinguishing the impacts of certification in terms of management of production processes and the business dimension.

3.1.1 The production process

The establishment of certification improves the internal production process of the company so that the service is actually produced reliably and efficiently [8]. Increased rigor established by quality standards is generally described as very positive constituting a “kind of ordering.” It eliminates the causes of errors and reduces the variability of the production process, therefore, become easier to control [8]. Thus, the use of these organizational tools leads to improved customer satisfaction [38], reduced costs and increased productivity [12].

Hypothesis 1. The quality assurance certificates increase productivity and reduce costs enabling to positively value added (VA).

3.1.2 The commercial process

A fierce national and international competitive pressure, the growing number of international outsourcing contracts, diversification of supply sources and production processes are just examples of the different factors that prompted the inception of assurance quality standards as a means of establishing trust between economic partners. In this context, [34] stressed that quality certification can be perceived as a differential signal into a commercial approach on the market. In a study based on 47 semi-structured individual interviews conducted in Quebec (Canada) between June 1998 and March 2001, [3] shows that half of the corporate employees (managers, quality managers and operators) surveyed tended to associate these standards with “commercial certificates” that are more suited to marketing purposes. Similarly, [33] show that 90% of surveyed companies justify the mobilization of the certification by the need to maintain a competitive edge and promote sales. They note that “procure a certificate and it is your company business card” thus granting enhanced customer satisfaction and easier access to new markets. Hence the hypothesis chosen in this study.

Hypothesis 2. The quality assurance certificates enhance the commercial aspect measured by turnover (CA).

3.2 The financial dimension

The implementation of quality assurance certification provides a guarantee to the markets that the products offered are manufactured and tested according to a minimum set of requirements intended to meet customer needs. It is also seen as an information signal on the stock market, sensitive to communication actions undertaken by companies [33]. In theory, including agency theory, markets should therefore respond positively to the opportunity offered to “prove” the quality of their production [34]. It is perceived as a signal to address the information asymmetry between the seller and buyer. This is validated in an empirical study by [2] on French companies in which the author shows certification translates into increased financial efficiency of 35% immediately after its installation. Similarly, [12] show, based on 21.482 certifications in the United States, that certification leads to a considerable improvement in financial performance. Also in the same report, [37] demonstrates that certification has positive impacts on financial returns for certified companies listed in Singapore. Therefore, this hypothesis is to be taken into account in this research paper.

Hypothesis 3. The quality assurance certificates have a positive impact on the financial aspect measured by the productivity of the financial capital.

3.3 The organizational dimension

Besides the operational and financial dimensions, business executives take interest in certification as being a relevant internal management tool. The new management literature shows that these standards can be considered potential catalysts for organizational change [1]. Proponents of certification maintain that this new management modality, via procedures and work standards, leads to a better homogenization of working methods and thus reduces the uncertainty of coordination and relational complexity. According to them, it transforms the contents and forms of knowledge as well as the relationships between the actors of production [8]. The prescription of quality standards helps to bring more rationality in the coordination of management, collection and knowledge creation. Furthermore, quality standards promote the codification of knowledge, creation and transfer as well as they lead to a clarification of knowledge about the organization and their formalization. In the same vein, the management literature shows that quality standards may be organizational learning devices [28]. They are seen also as sources of motivation because they relate to concerns that transcend the boundaries of the company and the strictly economic objectives that enhance the pride and employee engagement. They are therefore seen as an integrated technique as they tend to mobilize and strengthen internal cohesion around a common project company [1]. Hence the fourth research hypothesis:

Hypothesis 4. Quality assurance certificates have a positive impact on the internal management, which increases labor productivity (PRT).

4 Impact evaluation methods for the implementation of quality assurance standards on the performance of French companies

Due mainly to the aforementioned reasons and while taking into account the selection bias, the Rubin Model has been implemented in this work to test the causality between the use of quality assurance and safety standards on the one hand and the corporate performance on the other hand.

4.1 The Rubin Causal Model (1974)

Let us consider a set of firms i = 1, 2....... N making or not the object of a treatment (here implementing a quality and safety assurance program) andy_i denoting business performance variable (here added value, turnover, labor productivity and productivity of the financial capital). The Rubin Causal Model rests mainly on the assumption of two unobserved result variables namely Y_0i and Y_1i. These two variables are random variables associated with each company, and correspond to potential achievements of the performance variable depending on whether the company is certified or not. Potential Y_0i and Y_1i do exist before undergoing the treatment, but only one of them is observed. Thus, for a certified firm, Y_1i is observed while Y_0i is unknown.

A straightforward way to gauge the impact of quality certification is to consider the difference in mean performances of certified and non-certified companies C. This method rests on the assumption of independence between potential performance variables Y_0i and Y_1i and the variable of certification choice T_i = 1. In light of this assumption, it is to be noted that:

$\begin{matrix} C = E (Y_{1 i}) - E (Y_{0 i}) = E (Y_{1 i} / T = 1) \\ - E (Y_{0 i} / T = 0) \end{matrix}$ (1) The assumption of interdependence between variables of potential performances and variables of reorganization choice is very restrictive. Since the previous independence property is not satisfied, the natural estimator formed by the difference in mean outcome variables is affected by a selection bias:

$\begin{matrix} C = E (Y_{1 i} / T = 1) - E (Y_{0 i} / T = 0) \\ = E (Y_{1 i} / T = 1) - E (Y_{0 i} / T = 0) \\ + [E (Y_{0 i} / T = 1) - E (Y_{0 i} / T = 1)] \\ = [E (Y_{1 i} / T = 1) - E (Y_{0 i} / T = 1)] \\ + [E (Y_{0 i} / T = 1) - E (Y_{0 i} / T = 0)] \\ = C_{T = 1} + β \end{matrix}$ (2) Selection bias is the term β = E (Y_0i/T = 1) - E (Y_0i/T = 0)in the above expression. This bias arises out of the fact that the mean situation of certified companies has not been the same as that of those untreated companies. This difficulty can nevertheless be circumvented only if for each firm, certified or not, one can observe the mean performance of non-certified companies is the same as that of certified and non- certified reorganized firms. This comes down to the condition of independence between the potential performance variable of non-certified companies Y_0i and the treatment choice T:

$Y_{0 i} ⊥ T \Rightarrow E (Y_{0 i} / T = 0) = E (Y_{0 i})$

However, there are good grounds to believe that this will not usually be the case. Several key differences exist between certified and non-certified firms: size, industry, strategy and firm structure, pressure from both customers and suppliers, economic performance, market trends ... etc.. Therefore, (unconditional) independence between latent result variables (Y_1i and Y_0i) and treatment assignment T is very unlikely. A less restrictive condition is to consider that there is a set of conditional observable variables X to which the property of independence between latent results and reorganization is verified. Generally speaking, this property requires a thorough consideration of a large number of conditioning variables. For this reason, matching on observable characteristics can be difficult to achieve in practice. This problem was solved by [35] who proposed an alternative method which consists of estimating for each certified and non-certified company, a probability to be a member of a class held by a counterfactual master based on its observable baseline characteristics X: the propensity score, as noted.

$P (X) = Pr (T = 1 / X)$

4.2 Illustration of different methods and estimation steps

To test the impact of certification on firm performance, there is a wide range of estimators based on assumptions which have varying degrees of rigidity and whose implementation presents variable degrees of complexity [26]. In the present work, the following estimates are to be put to the test: the naive estimator, the nearest neighbor estimator, the caliper or Radius estimator and finally the matching estimator with kernel function.

4.2.1 The naive estimate

The naive estimate is the simplest solution according to some authors [14]. It consists in the difference between the arithmetic averages of Y_1i and Y_0i.

4.2.2 The estimate by the nearest neighbor [Nearest Neighbor (NN)]

The Nearest Neighbor matching is probably one of the most straightforward matching procedures [17]. A company from the comparison group is chosen as a match for a treated company in terms of the closest propensity score and thus the matching of the M nearest neighbors(M ≻ 1)is achieved. In this case study, M = 1 without replacement has been used. The estimator of the average treatment effect of the introduction of certification of certified companies is written as follows: $C_{T = 1} = \frac{1}{N_{1}} \sum_{T_{i} = 1} [Y_{i}^{1} - \sum_{T_{j} = 0} {ω_{N_{0}} (i, j) Y_{j}}^{0}]$ (3)

4.2.3 Caliper or radius matching

In line with this method, a company from the control group is matched with another company from the certified group located inside the caliper (or maximum propensity score distance by which a match can be made) and it is the closest in terms of propensity score. In this research, a level of intensity (radius) 0.01 is imposed, which avoids the risk of poor matches and improves the quality of the match [6].

In many-to-one (Radius) caliper matching, the estimator of program impact may be written as: $C_{T = 1} = \frac{1}{N_{1}} \sum_{T_{i} = 1} [Y_{i}^{1} - \sum_{T_{j} = 0} {ω_{N_{0}} (i, j) Y_{j}}^{0}]$ (4)

4.2.4 The estimate by matching with kernel function

The kernel matching estimator proposed by [22] and [23] offers different estimates following the previous terminologies by [19] and [16]. For them, the kernel matching uses all cases in the control group to construct the counterfactual for each reorganized company. The final estimator of the certification impact, certification in the present condition is as follows [23]:

$\begin{matrix} \hat{E} [Y_{0} ⊥ P (x) = P (x_{i})] \\ = \sum_{j \in I_{0}} \frac{K [(P (x_{j}) - (P (x_{i})) / h]}{\sum_{j \in I_{\circ}} K [(P (x_{j}) - (P (x_{i})) / h]} y_{j} \end{matrix}$ (5)

4.3 The steps of the estimate

The estimation procedure involves three main steps: calculating the propensity score, determining the propensity score and finally the estimate to speak

4.3.1 Calculating the canonical score

In this stage, it seems convenient to estimate the probability of the certification choice for each company or the canonical score. The purpose of this estimate is to explain the T variable of the certification choice through the X observable characteristics. While assuming that this choice made by the company is only the visible manifestation of a latent variable leading to the certification choice, the logit model allows for testing whether the X observable characteristics suspected of influencing the decision of certification choice have a significant say in the implementation of a new quality assurance program. This step is informative inasmuch as it proposes a description of the treatment assignment.

4.3.2 Determining the propensity score

Once each company’s propensity scores were calculated, the common-support of the densities scores of the two groups of companies is determined. The question of determining the distribution support of the propensity score is essential in this type of analysis [23]. This support indicates the existence of both the density of the propensity score for both certified and non-certified firm groups.

4.3.3 The estimate to speak

Just apply the formulas.

5 Data presentation

Mobilized data are taken from the matching of the 2006 COI survey, jointly produced by DARES (Directorate for Research, Studies and Statistics) and INSEE (The National Institute of Statistics and Economic Studies), with those data related to businesses’ performance from DIANE files.

5.1 Presentation of the survey COI 2006

The survey conducted by the COI in 2006 is the result of a growing interest by researchers and international organizations, in the late 1980s, in the organizational and technological business dynamics. It is a deep extension of the first COI 1997 survey [20]. The survey was carried out simultaneously on a representative sample of industrial companies and a representative sample of their permanent employees. The development and coordination of investigative device were performed in the France’s Centre for Employment Studies and produced by many statistical agencies. The survey design was developed by a number of researchers from various disciplines. The survey questioned 16,962 companies with 10 or more employees divided into the following segments: sections D (Manufacturing), E (Production and distribution of electricity, water and gas), F (Construction), G (Trade), H (Hotels and restaurants), I (Transport and communication), J (financial intermediation), K (Real estate, renting and business activities) and groups 92.1 and 92.2 (motion picture and video activities, radio and television). Public companies in these sectors are part of the IOC field 2006. The overall response rate was 85%.

5.2 DIANE

DIANE files are composed of information available in the tax files, namely, in particular, the balance sheet and income statement of 885.422 French companies in 2011. These companies are divided into four targets: A (140.074 companies), B (139.985 companies), C (133.574 companies) and D (471.816 businesses). This source allows for identifying 113,823 manufacturing companies.

Information is available from the annual financial statements (balance sheet, income statement, statement of assets and liabilities...) that all public limited company (SA), limited liability company (SARL), a general partnership whose partners are SARL or SA and agricultural cooperatives whose turnover exceeds EUR (euro) 75,000, are required to file with the clerks of the commercial courts under penalty of a fine of 1500 euro [27]. DIANE files are used throughout this work in order to gauge the economic and financial performance of companies.

5.3 Matching data

As noted earlier, 16.962 companies of 10 or more employees responded to the 2006 COI survey. These companies spread across 8 fields of economic activity. In this case study, only the manufacturing industry has been included. Of the 3962 manufacturers documented in the file of the COI survey, 2339 were found in DIANE files for the years 2005, 2006 and 2009. Companies which showed negative added value as well as companies in which at least one of the growth rate of variables (value added, turnover, labor productivity or financial performance) shows an outlier, were then excluded from the data set [27]. In total, the selected sample consists of 2106 manufacturing companies with 10 employees or more in 2006, only 53.1% of the initial manufacturing companies were kept.

6 Results and comments

6.1 Logistic regression and construction of common support

The estimation results of the logit model presented in Table 2 suggest a significant link between the introduction of certain organizational innovations and the implementation of quality certification. As a matter of fact, results show that belonging to a group as well as being part of a network and the use of a traceability tool play a major role in the certification probability. However, the implementation of a logistical tool that leads to a low probability of certification.

Also in the same vein, it has been noted that the use of ICTs, such as the Web, the Intranet and the Extranet, increases significantly the likelihood of implementing certification. In terms of relations with customers and suppliers, results indicate that the requirement of first tier customers to comply with quality standards or with quality control procedure and the provision for a data-processing program accessible to the major customers significantly increases certification probability. Conversely, the probability of being certified proves to be lower when first tier suppliers of the firm have access to a paired data-processing program.

From the outcome of the estimate of logistic regression, propensity scores of each certified and not certified firm have been established while holding only the significance variables of 10% or lower. Figure 1 shows that the surface of the common propensity score covers a very large distance: the common support region obtained is more or less [0.18, 0.95].

6.2 Results and interpretation of the estimated impact of certification on corporate performance

Once the propensity score and its common support were determined, one can assess the impact of certification of businesses through the growth rate of the added value, the turnover, labor productivity and profitability of financial capital between 2006 and 2009 through the use of the following three methods: the estimation by the nearest neighbor (1), the radius or caduis estimate (h = 0.01) and, last of all, the weighted estimate. The outcomes of these tests are presented in the following table:

Initial findings of the effect of certification on corporate performances indicate, as shown in Table 3, that naïve estimates differ from those produced by other estimates.

In the first place, it appears that naïve estimator and the kernel estimator reveal a positive, insignificant effect certification on growth rates of added value. Conversely, the Nearest Neighbor estimator of distance 1 and the radius estimator show a negative and insignificant correlation between implementing assurance quality certification and the growth rate of the added value.

Furthermore, with the exception of the kernel estimator matching, naïve estimator, nearest neighbor estimator and Radius estimator indicate a positive insignificant impact on the treatment effect on the growth rate of the turnover.

What makes the Naïve estimator stand out from the other estimators is mainly the growth rate of labor productivity. As a matter of fact, the global effect of implementing quality certification on the growth rate of productivity is positive and significant to 6.6%. This causality effect remains positive with the use of the Kernel estimator but loses its significance while it is both negative and insignificant by employing the Nearest Neighbor estimator and the Radius estimator.

Finally, it should be noted that the global effect of the implementing quality certification on the growth rate of the financial capital productivity is positive and significant in all estimators. This effect appears to be strong and quite significant in light of the naïve estimator: certified companies reported an improved growth rate of 10.8% in their financial capital productivity in comparison with non-certified companies.

To sum it up, the results of this research paper seem to be partially in tune with the proposed hypotheses. They indicate that the inception of quality certificates leads to positive effects on the productivity of the financial capital. Our results are consistent with the findings of the financial approach which considers the implementation of quality and assurance certification a boost to the financial profitability. However, results show that the inception of quality certificates do not have any positive effect on the added value, turnover and the labor productivity. Though the results obtained may present some discrepancies with the hypotheses in use, that should not come as a surprise since they tally with those obtained by [15], [29] and [30].

They align with the analyses proposed by the theory of job sociology [8, 32]. They blame the cost of certification and the cumbersome nature of procedures. [4] signals that the advocators of this approach argue that quality standards, while requiring a compilation of documentation designed to record procedures and ensure management and strict monitoring of quality, can be conducive to an excessive bureaucratization. They also can be seen as a coercive control tool and thus cause the eruption in the “iron cage”, to use Weber’s words [40, 41]. The introduction of this mechanism can bring us back to the standard Taylor model [39], as pointed out by [11].

7 Conclusion

The aim of the present study is to evaluate the treatment effect of implementing assurance and quality certifications on the economic performances of 2106 french manufacturing companies.

To this end, the econometric matching algorithm as inspired from Rubin Causal Model (1974) has been employed along with some recent estimation methods covered in this study. These methods furthered by some econometricians take pride in proposing new non parametric estimation techniques that allow for the evaluation of the effect of a quality assurance program on both certified and non certified businesses. They can also provide efficient and convergent estimators while taking into account the selection bias. In this study, the methods in use range from the Naïve estimator to the Nearest Neighbor estimator (M = 1) to the caliper estimator (h = 0.01) and last the Kernel method.

Results show that while the inception of the quality assurance certification has improved the productivity of the financial capital, it has no effect on the added value, turnover and labor productivity of businesses.

As with all such work, the present paper does present certain limitations. The first one can be ascribed to the choice of the performance indicators since the performance of the businesses has been narrowed to one and only dimension: the economic dimension. However, since the 1990’s, the concept of performance has become multidimensional [24]. As a consequence, it would be more appropriate to evaluate the certification/performance while taking into consideration other levels of analysis.

The second limitation has to do with the source of data use which covered the period up to 2006. This study suffers from a limited time interval during which the relationship between the implementation of the certification program and performance is tested. A growing number of authors consider that, unlike longitudinal studies, cross sectional studies cannot offer a dynamic vision. They provide less informative data, less variability and more collinearity between two variables [21].

References

Berger-Douce

, La certification ISO 14001, catalyseur du changement organisationnel? L’expérience de deux maisons de champagne. Communication á la 11^ème Conférence de l’AIMS. Paris; 2002.

Bernardi

, Certification et performance boursière. Revue Sciences de Gestion; 2001.

Boiral

, ISO 9000: De l’impératif commercial au rite de passage organisationnel. Communication à la 11^ème Conférence de l’AIMS. Paris; 2002.

Boiral

, Mettre en øeuvre ISO 14001: De la quête de légitimité à l’émergence d’un mythe rationnel. Communication à la 13^ème Conférence de l’AISM. Le Havre; 2004.

Boiral

, Mongour

, Les impacts d’ISO9000 sur l’amélioration des performances: Une revue systématique. Communication à la 18^ème Conférence de l’AIMS. Grenoble: France; 2009.

Caliendo

and Kopeinig

, Some Practical Guidance for the Implementation of Propensity Score Matching. Journal of Economic Surveys 22(1) (2005), 31–72.

Campinos-Dubernet

and Jougleux

, L’assurance qualité: Quelles contributions à la qualité des services? Revue Française de Gestion 5(146) (2003), 81–98.

Campinos-Dubernet

and Marquette

, Une rationalisation sans norme organisationnelle: La certification ISO 9000. Sciences de la Société 46 (1999), 83–101.

Canard

and Boronat

, Mangement par la qualité totale et changement organisationnel, M. dans Rainelli, J.L. Gaffard, A. Asquin. Les nouvelles formes organisationnelles. Economica 217–234.

10.

Cherfi

, La qualité: Démarche, méthodes et outils. Lavoisier, 2002.

11.

Cochoy

, Garel

J.P.

and De Terssac

, Comment l’écrit travaille l’organisation: Le cas des normes ISO 9000. Revue Française de Sociologie 39(4) (1998), 673–699.

12.

Corbett

C.J.

, Montes-Sancho

M.J.

and Kirsch

D.A.

, The Financial Impact of ISO 9000 Certification in the US: An Empirical Analysis. Management Science 51(7) (2005), 1046–1059.

13.

Dubois

, Doutre

and Koubenan

R.D.

, Participation dans les démarches qualités et confiance sociale. Science de la Société 1999.

14.

Duguet

, Les subventions à la recherche et développement sont-elles substituables ou complémentaires à leur financement privé? Une étude micro économétrique. Document de Travail. Université de Paris I. Cahiers de la MSE; 2002.

15.

Easton

G.S.

and Jarrell

S.L.

, The Effects of Total Quality Management on Corporate Performance: An Empirical Investigation. Journal of Business 71(2) (1998), 253–307.

16.

Fougère

, Les méthodes économétriques d’évaluation. Revue Française des Affaires Sociales 1-2(12) (2010), 105–128.

17.

Givord

, Méthodes économétriques pour l’évaluation de politiques publiques. Document de travail. Direction des Études et Synthèses Économiques; 2010.

18.

Gorgeu

and Mathieu

, L’assurance qualité fournisseur de l’industrie automobile française. Revue d’ Economie Industrielle 75(75) (1996), 223–237.

19.

Goussé

, Murard

, Naouas

and Silavong

, Les estimateurs de matching: Une alternative à l’expérimentation aléatoire ? D’après l’étude de B. Crépon, L. Behaghel et M. Gurgand. ENSAT: Paris; 2010.

20.

Greenan

, Guillemot

and Kocoglu

, Annexes. Ré Seaux 4(162) (2010), 275–304.

21.

Gujarati

D.N.

, Économétrie. De Boeck & Larcier; 2004.

22.

Heckman

J.J.

, Ichimura

and Todd

P.E.

, Matching as an Econometric Evaluation Estimator: Evidence from Evaluating a Job Training Programme. Review of Economic Studies 64(4) (1997), 605–654.

23.

Heckman

J.J.

, Ichimura

and Todd

P.E.

, Matching as an Econometric Evaluation Estimator. Review of Economic Studies 65(2) (1998), 261–294.

24.

Hounkou

, Les pratiques de gestion des ressources humaines et les performances des entreprises béninoises: Une analyse par la méthode de corrélation canonique. Revue Internationale sur le Travail et la Société 9(1) (2011), 17–40.

25.

Igalens

, La qualité, dans Allouche J, Encyclopédie des ressources humaines. Vuibert: Paris; 2006.

26.

Janod

and Saint-Martin

, Impact des Réorganisations du Travail sur les Performances des Entreprises Manufacturières: Une Analyse sur Données Françaises 1997-1999. Revue Économique 54(3) (2003), 637–648.

27.

Janod

, Changement Organisationnel, Qualifications et Croissance. Thèse de Doctorat en Science Economiques. Université de Paris 1: Pantheon-Sorbonne; 2002.

28.

Lambert

and Loos-Baroin

, Certification ISO 9000 et création de connaissances opérationnelles ou conceptuelles: Une étude de cas. Finance-Contrôle- Straté gie 7(1) (2004), 53–79.

29.

Martinez-Costa

and Martinez-Lorente

, A triple analysis of ISO 9000 effects on company performance. Journal of Productivity and Performance Management 56(5) (2007), 484–499.

30.

McGuire

S.J.

and Dilts

D.M.

, The financial Impact of Standard Stringency: An Event Study of Successive Generations of the ISO 9000 Standard. International Journal of Production Economics 113(1) (2008), 3–22.

31.

Reix

, Savoir tacite et savoir formalisé dans l’entreprise. Revue Française de Gestion; 1995.

32.

Reverdy

, Les formats de la gestion des rejets industriels: Instrumentation de la coordination et enrôlement dans une gestion transversale. Sociologie du Travail 42(2) (2000), 225–243.

33.

Rolland

and Tran

, La certification qualité est-elle un facteur de compétitivité pour les entreprises? Le cas de la norme ISO 9001. Acte de la 5^ème édition du colloque métamorphose des Organisations Nouvelles régulations. Nancy; 2006.

34.

Rolland

, Un bilan de 20 ans de certification des systèmes de management de la qualité: Les apports perçus de la certification ISO 9000 par les managers. Management & Avenir 9(29) (2009), 31–51.

35.

Rosenbaum

P.R.

and Rubin

D.B.

, The central role of the propensity score in observational studies for causal effects. Biometrika 70(1) (1983), 41–55.

36.

Rubin

, Estimating causal effects of treatments in randomized and non- randomized studies. Journal of Educational Psychology 66(5) (1974), 688–701.

37.

Sharma

D.S.

, The Association between ISO 9000 and Financial Performance. The International Journal of Accounting 40(2) (2005), 151–172.

38.

Singh

P.J.

, Empirical Assessment of ISO 9000 Related Management Practices and Performance Relationships. International Journal of Production Economics 113(1) (2008), 40–59.

39.

Taylor

, Principes d’organisation scientifique des usines. Dunod: Paris; 1912.

40.

Weber

, Economy and Society: An outline of interpretive sociology. New York: Bedminister; 1968.

41.

Weber

, The protestant ethic and the spirit of capitalism. New York: Scribners; 1958.