INFORMATION TECHNOLOGY OUTSOURCING IN FRANCE AND GERMANY Dominique Geyer Groupe ESC Nantes Atlantique n°72 - juin 1999

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

Information Technology Outsourcing in France and Germany

1

Abstract The aim of this article is to analyze the determinants of a new phenomenon, outsourcing of whole or part of the information systems (IS) function of a company (or ITO : Information Technology Outsourcing), in two countries: France and Germany. It highlights the influence of institutional and cultural factors. A postal survey carried out with 160 large German and French companies underlines differences in the ITO practices of the firms of these two countries. Forty-five percent of the firms of the sample carried out (or are about to carry out) an ITO. Results indicate that German companies do not outsource more frequently than do French companies, but they tend to accompany the ITO by holding capital of the vendor. Moreover, the ITOs with reduction of personnel are more numerous in France than in Germany.

Résumé L’objectif de cet article est d’analyser les déterminants d’un nouveau phénomène : l’externalisation de tout ou partie de la fonction informatique de l’entreprise (ou infogérance). Cette pratique est étudiée en France et en Allemagne car il s’agit de déterminer l’influence de facteurs culturels et institutionnels. Une enquête postale menée auprès de 160 grandes firmes françaises et allemandes a décelé des différences dans leur comportement d’externalisation. 45% des firmes de l’échantillon ont externalisé (ou sont sur le point de le faire). Les firmes allemandes n’externalisent pas plus souvent que les firmes françaises, mais elles ont tendance à accompagner l’opération d’infogérance d’une prise de participation dans le capital du prestataire informatique. De plus, les infogérances comportant des réductions de personnel sont plus nombreuses en France qu’en Allemagne.

Mots-clés - Key-words Externalisation Fonction informatique Facteurs culturels France Allemagne

Outsourcing Information system Cultural factors France Germany

© Groupe ESC Nantes Atlantique - Cahiers de Recherche du CREA - 1999 Toute reproduction de ce document est interdite sans l’autorisation expresse de son auteur 1

I am grateful to David Curry for extensive comments on an early draft of this paper. 2

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

1. Introduction

If the concept of firm is represented as a circle, one can distinguish two types of movement: integration and outsourcing. Whereas corporate or functional integration is the subject of many articles, outsourcing has been relatively neglected by researchers. However, at the present time, many firms have exceeded a critical size. One thus witnesses an increased need for outsourcing. But is outsourcing simply a passing fad or is it a permanent fixture of business in the 21 st century ? The answer to this question is not simple because firm boundaries often evolve sporadically. Initially, the firm develops and grows very quickly. In a second phase, a wave of outsourcing counterbalances the rapid growth and corrects certain problems that have arisen internally. The empirical framework selected for the present study views outsourcing as whole or part of the IS function of companies in two countries: France and Germany. The objective is to analyze the determinants of ITO behavior. What factors are likely to influence the ITO behavior of the firm?

To highlight the separate and interactive affects of cultural and

institutional factors, firms from both cultures are compared. The comparison between France and Germany is particularly interesting due to important differences in management practices as well and institutional environments between the two countries.

In the first section of the paper, we define Information Technology Outsourcing. We highlight various forms of ITO and characterize outsourcing as a governance structure. Section two highlights the importance of institutional factors.

Section three states five

hypotheses to be tested empirically. Section four then presents results from the survey. The paper concludes with a summary of the contributions of the present research. 3

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

2. ITO: definition, forms and characteristics

2.1 Definition

Outsourcing whole or part of the IS function means to entrust to a third company (the outsourcer or vendor) physical and/or human resources of whole or part of the IT infrastructure. Outsourcing can be temporary or permanent. The company-customer may or may not participate in the capital of the vendor. This definition of the ITO needs additional refinement. ITO is a new phenomenon, an innovation. It acts within the meaning of D. Teece [ 1980] as an administrative innovation (as opposed to a technological innovation). One could even speak about management innovation insofar as it implies significant changes in internal alignments as well as external alignments. Typically there is a change in governance : from a hierarchical governance structure, possibly supplemented by a market governance structure, to a hybrid governance structure as defined by Williamson [ 1991 ].

2.2 Forms of ITO

This section is devoted to clarifying various forms of

ITO.

Note first that

outsourcing cannot be defined by the simple dichotomy : to make everything or to buy everything. ITO modifies but cannot obliterate the role of IS management. Controlling an internal team is very different from controlling an external agent (the outsourcer). At present, ITO has the following forms : 1 - IS function ;

4

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

2 - Development of applications: the creation of software that will be adapted to the needs of the company; 3 - Networks: at the corporate level, a network makes it possible to exchange files, access the same programs, share applications , etc. 4 - Design of system: this is the operation of reorganizing to improve or reconfigure the IS of the company (Examples : downsizing, moving from mainframe computers to client/server systems) ; 5 - Integration of system: this operation includes integrating materials and software standards so that the system conforms to user needs and requirements ; 6 - Microcomputers ; 7 - Data centers.

3. Research hypotheses

In order to explore the determinants of the ITO behavior of the German and French companies, we considered three levels of analysis: the IS function,

the firm and the

institutional environment.

3.1. Characteristics of the IS function

The characteristics of the IS function - its nature as a responsibility center and its degree of centralization - can influence the ITO behavior of the firm. The IS function of the company is a responsibility center. A responsibility center is an organization unit that is supervised by a manager who is responsible for its activities. A company is a collection of responsability centers. The IT manager must achieve certain results during a predefined 5

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

planning horizon using the means at his disposal.

There are two general types of

responsibility centers : cost centers and profit centers. In the late 50' s when IT was first introduced, management didn' t know how to control this new resource. For a long time, the control of IT management was minimal. Initially, the IS function was a cost center . In a cost center, control concerns only costs . In a profit center, control concerns costs and revenues. IT particularly prior to the 80' s was not subject to strenuous monitoring. The specificity of IT resources allowed the function to grow without major constraints on IT budgets. However, in many companies, the durable crisis caused a change of status from cost center to profit center. The IS function is today held responsible for at least a portion of the firm' s profitability. Today, the IT function receives an income statement so that it can measure its contribution to corporate profit. The transition from a cost center to a profit center allowed better control of costs.

The choice of outsourcing answers the same objective. This argument leads to

hypothesis one : H1: Companies that manage their IS function as a profit center outsource all or part of their IS function more often than do other companies.

A second characteristic of the IS function that is likely to have an influence on ITO behavior is the degree of centralization. Even more than most corporate functions, the IS function is not an island insulated within the company because IT is involved in virtually all of the other functions of the company. Despite this growing connectivity, IS is still an autonomous department in most companies (centralized IS function).

This degree of

centralization can be better appreciated by recalling how IS is typically represented in corporate organizational charts.

Often, IS is shown as a staff function to finance and

accounting reflecting the first applications of data processing in these fields.

6

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

The proliferation IT applications and the push to involve users in IT design processes have without a doubt contributed to the decentralization of the IS function. This tendency has been amplified by recent developments of new IT. Whereas the boundary of a centralized IS function is easy to determine within a company, identifying this boundary for a decentralized IS function is much more difficult. Decentralization can be defined using various criteria : geographical, technical, etc. The arguments in favor of decentralization constitute an answer to the disadvantages of centralization, e.g. reduction of reaction times, narrower relations with users, increased specialization of computer staff with respect to user needs. In reality, the two idealized forms of the IS function centralized vs decentralized, are seldom seen in practice.

For this reason, we introduce a third type: a hybrid. How can the degree of

centralization - decentralization of the IS function of the company influence its ITO behavior ? When the IS function is centralized, the bonds with the other functions of the company are more reduced than with a decentralized IS function. Outsourcing means to break on set of bonds in order to create another set. For the centralized IS function, the bonds to be broken are fewer.

The transition costs thus will be lower

for the centralized than for the

decentralized IS function. The second hypothesis is : H2: Companies with centralized IS function's outsource all or part of their IS function more often than do other companies.

3.2. Characteristics of the firm

Do companies having a certain type of structure have a higher tendency to ITO? IT is involved in all functions of the company. However, modern multidivisional firms (or M-form firms) are typically divided into a number of autonomous units, e.g. independent business units . The bonds between these various units seem much fewer in number and weaker than 7

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

the bonds in a functional structure. In contrast to the M-form firm, the functionally organized firm (or U-form firm) is a firm in which the functions, such as manufacturing, sales, accounting, and so on, are each controlled by a single department. Thus separating IT from an M-form firm will be easier to carry out than separating IT from a U-form firm. The transition costs are lower. The third hypothesis therefore is : H3: Companies having a divisional structure outsource whole or part of their IS function more often than the companies having another structure.

3.3 Characteristics of the institutional environment

Hypotheses H4 and H5 require knowledge about differences in ITO practices between German firms and French firms. Differences are found on two different levels : in the respective legal environments and in the economic systems of the two countries. Regarding the national economies, we seek specific aspects in each economy that influence ITO practices of firms headguartered in each country. For example, trade unions play vastly different roles in France and Germany. Among the OECD countries, France has the lowest unionization rate. French trade unions are many but they are not very powerful. On the other hand,

German trade unions are few but powerful. In Germany, the D.G.B. (Deutsche

Gewerkschaftsbund) has a hegemonic position.

The German trade-union system has a

cooperative relationship with German firms. French trade unions are governed by conflict logic rather than cooperative logic. The German trade-union system is a dual system ; i.e. the work' s council cooperates with management. It is the trade union, that is not represented in the company, which decides to strike. In France, the situation is different because there are trade-union sections (sections syndicales) in each company. Representation is guaranteed by the Matignon agreements of 8

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

June 1968 in companies having at least 50 employees. Thus, there are important differences on the level of the trade-union traditions in France and Germany. Relations between employer and employees are more cooperative in Germany than in France ; Germany differs from other European countries because it uses an original system of comanagement (Mitbestimmung). Since 1870, German corporate law requires that a joint stock company (Aktiengesellschaft) create two distinct bodies:

a supervisory board

(Aufsichtsrat) and a board of directors (Vorstand). In Germany, the system of comanagement works perfectly. The German law of 1976 on comanagement imposes an employee profit sharing on the inspection committee according to a rigorously defined system. The employee representation quota varies according to the type of company. Thus, in the iron, steel and mining industry, the number of employee representatives equals the numbers of shareholder representatives. However, for other sectors, employee representation constitutes one third of the members of the board of directors. How might corporate culture differences influence the ITO behaviors of the German and French firms? An ITO operation affects both physical and human resources in a firm. Outsourcing with personnel changes (dismissal, transfer of employees to the vendor) evoke hostile reactions from employees. However, the decision of whether or not to outsource is passed down from upper management and/or from the IS director. When superiors impose outsourcing on rank and file employees, they must consider legal and cultural factors. Various legal provisions are formal constraints according to D.C. North [1990]. We classify these formal constraints as external because they originate outside the company. These constraints appear to be much stronger in Germany than in France, i.e. German employees can more effectively oppose outsourcing that

leads to reductions of personnel than can French

employees due to the greater power of German trade unions and the German comanagement system. Our fourth hypothesis therefore is : 9

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

H4: Personnel changes are more likely to accompany outsourcing decisions in France than in Germany.

Do German and French firms have different characteristics ? Do these differences influence the practices of outsourcing ? The German economy is characterized by the presence of industrial groups such as the Konzern which can be compared to the Japanese Zaibatsus. Similar groups can also be found in France but these groups were formed primarly in reaction to French industrial policy and, for a long time, were public. The public sector in Germany is less important ; good industrial policy in Germany is basically no policy. In France, until the mid-seventies,

public authorities tried to create the best conditions for external

competitiveness on world markets in selected industries, primilarly to compete with German and US Firms. This was the period of ' national champions'(Renault, Thompson, CGE, etc), during which increased size of operation was the discriminating factor for efficiency. In Germany, the Konzerns do not owe their formation to public stimulants. Rather their emergence and growth stems from German economic power. Germany stands apart from other European countries due to the presence of Konzern and their tightly relations. For example, the bonds between banks and manufacturers are much closer in Germany than in France where the banking structure is weaker, more fragmented, and works less closely with industry. How do the characteristics of the German economy (presence of Konzerns) induce differences in the ITO behavior between German firms and their French counterparts ? In fact, the Konzerns often create their own subsidiary companies and entrust management of the IS function to them. These subsidiary companies can offer their services to all of the firms of the Konzern. The larger the Konzern, the more subsidiary companies are created in order to profit from economies of scale. So the fifth and last hypothesis is : 10

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

H5: German companies use quasi-outsourcing in the IT field more often than the French companies.

4. Results of the empirical study

4.1 The sample

To test the five research hypotheses, we sent a native language version questionnaire to the largest German and French companies. This questionnaire was constructed with the help of three IT consultants (two French and one German). The questionnaire was pretested with five French companies and seven German companies and reformulated. The size of a company is indexed by the company' s turnover defined as total assessment for banks and total premiums for insurance companies for the year 1993. The year under review selected is 1993. The questionnaire was mailed in October 1995 to the 500 largest firms in France and Germany. These firms represent various stages of the ITO decision-making process. The questionnaire contained filters to distinguish four categories2 of firms : Category 1 - Firms that have not considered outsourcing all or part of their IS function. Category 2 - Firms that actively examine the question of outsourcing all or part of their IS function but that had not yet finalized the decision at the time of data collection. Category 3 - Firms that had made a positive decision regarding outsourcing. Category 4 - Firms that had made a negative decision as regards outsourcing. Identifying who to target within each firm was difficult because outsourcing decisions can be made in many ways. However, trade articles reveal two primary groups responsible for this

2

Usually, the empirical studies devoted to the alternative to make-or-buy comprise only two categories: internal management or external management. The four categories suggested are closer to reality. 11

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

decisions ; general management and IS management. The decision can also be made jointly by these two parties. In the present study we addressed the questionnaires to the manager of IS in each firm.

4.2. Test of hypotheses

One thousand questionnaires were mailed and one hundred and sixty useable were returned : 12.2 % of the French mailing and 19.8 % of the German mailing. Table 1 gives a breakdown of returns by category and country. A Chi-Square test reveals no overall pattern of returns by category by country , i.e. the sample is relatively well-balanced in the 8 cells.

Table 1: ITO Behavior and nationality of the firms Germany

Chi-square

France Total 16 12 28 16.2 % 19.7 % 17.5 % 15 4 19 15.1 % 6.6 % 11.9 % 43 29 72 43.4 % 47.5 % 45.0 % 25 16 41 25.3 % 26.2 % 25.6 % 99 61 160 100.0 % 100.0 % 100.0 % Degrees of freedom Significance Contingency coefficient

2.76886

3

Category 1 Category 2 Category 3 Category 4 Total

0.42865

0.13043

Twenty-eight companies (17.5 %) of the sample had not yet considered the question of ITO (Category 1). Nineteen companies (11.9 %) had examined the appropriateness of ITO, but no decision had yet been made in these cases (Category 2). One hundred and thirteen companies (70.6 %) had studied the matter of ITO appropriateness and eighty-two companies

12

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

(45.0 %) had actually outsourced (Category 3) and forty-one companies (25.6 %) had considered but did not go forward with outsourcing (Category 4). Table 2 gives more detail about the precise nature of the IS function among firms in our sample.

Table 2: ITO forms in France and Germany Number of firms IS function

France 29

Germany 43

Total 72

Chi-square

13 14 27 33.4 % 19.4 % 24.4 % 9 9 18 23.1 % 12.5 % 16.2 % 2 11 13 5.1 % 15.3 % 11.7 % 3 7 10 7.7 % 9.7 % 9.0 % 2 4 6 5.1 % 5.5 % 5.4 % 3 19 22 7.7 % 26.4 % 19.8 % 5 2 7 12.8 % 2.8 % 6.3 % 2 6 8 5.1 % 8.4 % 7.2 % 39 72 111 100.0 % 100.0 % 100.0 % Degrees of freedom Significance Contingency coefficient

14.96880

7

Development of applications Networks Design of system Integration of system Data centers Microcomputers Others Total

0.03640

0.34472

Table indicates that the most outsourced components are the IS function (24.4 %) and data centers (19.8 %).

13

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

4.2.1. Influence of the characteristics of the IS function

The five stated hypotheses will be checked using Chi-square tests applied to the following five contengency tables.

Table 3: ITO Behavior and status of the IS function as a responsibility center Profit center

Chi-square

Cost center Total 9 16 25 12.5 % 20.8 % 16.8 % 5 13 18 6.9 % 16.9 % 12.1 % 41 28 69 56.9 % 36.3 % 46.3 % 17 20 37 23.6 % 26.0 % 24.8 % 72 77 149 100.0 % 100.0 % 100.0 % Degrees of freedom Significance Contingency coefficient

8.04935

3

Category 1 Caregory 2 Category 3 Category 4 Total

0.04500

0.22639

We note that χ² is significant at the α=0.05 level indicating that there is dependence between the ITO behavior of the firm and the status of the IS function as a responsibility center. Note that 56.9 % of the firms managing their IS function as a profit center use outsourcing in whole or part. For the companies managing their IS function as a cost center, this percentage falls to 36.3 %. These data thus support hypothesis one.

14

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

Table 4: ITO Behavior and degree of centralization of IS function Centralized IS function Category 1

Hybrid IS function

Decentralized IS function

Total

Chi-square

12 15.6 % 10 13.0 % 36 46.7 % 19 24.7 77 100.0 % Degrees of freedom

12 2 26 19.7 % 13.4 % 17.0 % 8 0 18 13.1 % 0.0 % 11.8 % 26 8 70 42.6 % 53.3 % 45.7 % 15 5 39 24.6 % 33.3 % 25.5 % 61 15 153 100.0 % 100.0 % 100.0 % Significance Contingency coefficient

3.16006

6

0.78850

Category 2 Category 3 Category 4 Total

0.14225

Table 4 shows that χ² is not significant at the α=0.05 level indicating that a firm' s ITO behavior and the degree of centralization of the IS function are not significantly statistically dependent. Thus the second hypothesis must be rejected.

4.2.2. Influence of the characteristics of the firm Table 5: ITO Behavior and structures of company Divisional structure Category 1

Matrix structure

Functional structure

Chi-square

9 13.5 % 10 14.9 % 36 53.7 % 12 17.9 % 67 100.0 % Degrees of freedom

1 5.5 % 3 16.7 % 9 50.0 % 5 27.8 % 18 100.0 % Significance

13.51790

6

0.03551

Category 2 Category 3 Category 4 Total

15

Total

18 28 26.9 % 18.4 % 5 18 7.5 % 11.8 % 22 67 32.8 % 44.1 % 22 39 32.8 % 25.7 % 67 152 100.0 % 100.0 % Contigency coefficient 0.28578

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

Table 5 shows χ² is significant at α< 0.03 which indicates dependence between the ITO behavior and the structure of the firm. In the sample, 53.7 % of the companies having a divisional structure outsourced whole or part of their IS function. For the companies having a functional structure, this percentage falls to 32.8 %. Thus, hypothesis 3 is supported.

4.2.3. Influence of the characteristics of the institutional environment

Table 6: ITOs with or without reduction of personnel France ITO with reduction of personnel ITO without reduction of personnel Total

Germany

Total

Chi-square

20 18 38 69.0 % 41.9 % 52.8 % 9 25 34 31.0 % 58.1 % 47.2 % 29 43 72 100.0 % 100.0 % 100.0 % Degree of freedom Significance Contingency coefficient

5.10548

1

0.02385

4.07584

1

0.04350 with Yates correction

0.25732

In the sample, 52.8 % of the ITO' s involved a reduction of personnel. However, this percentage masks the existence of a strong disparity between the German firms and the French firms. The chi-square test reveals that downsizing differs by country. For the French firms, 69.0 % of the operations of ITO lead to workforce reductions. This percentage is only 41.9 % for the German firms. Therefore, hypothesis 4 is strongly supported by these data. However, the degree of downsizing is relatively small ; e.g. in 36.8 % of the cases, the number of people concerned is less than ten.

16

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

Table 7: Quasi-outsourcings in France and Germany France

Chi-square

Germany Total 5 20 25 17.2 % 46.5 % 34.7 % 24 23 47 82.8 % 53.5 % 65.3 % 29 43 72 100.0 % 100.0 % 100.0 % Degrees of freedom Significance Contingency coefficient

6.54657

1

0.01050

5.31888

1

0.02109 with Yates correction

Quasi-outsourcing Outsourcing Total

In Germany,

0.28870

outsourcing without stockholdings are more numerous than

outsourcings with stockholdings (46.5 % of the German ITO operations and 17.2 % of the French ITO operations). The Chi-square test (with Yates correction) underscores a dependence between these two variables. Hypothesis 5 is, therefore, supported. An examination of the degree of subsidiarization reveals that only a few companies (two cases in France and eight cases in Germany) use subsidiaries to supply outsourced IT functions.

5. Conclusion

The objective of this study was to understand the ITO behavior of large French and German firms. The phenomenon of ITO was described as an organizational innovation. This practice is comparable with the hybrid governance structure developed by O.E. Williamson. The determinants of ITO were analyzed on three levels: the IS function level, the firm level and the institutional environment level. Quantitative analysis of a mail survey carried out with large French and German firms showed how factors like the status of the IS function as a responsibility center, the type of structure and the institutional characteristics, affect the decision to outsource IT functions.

17

Les Cahiers de Recherche du CREA - n°72 - Dominique GEYER - juin 1999

References

Coase, R.H. [1937], The nature of the firm, Economica, 4, 386-405 North, D.C. [1990], Institutions, institutional change and economic performance, Cambridge University Press Perry M. K. [1989], "Vertical integration : determinants and effects," pp. 183-255 in : R. Schmalensee and R.D. Willig (eds.), Handbook of Industrial Organization, Elsevier Science Publishing : Amsterdam. Teece, D. [1980], The diffusion of an administrative innovation, Management Science, 26, 464-470 Williamson, O.E. [1989], "Transaction Costs Economics", pp. 135-182 in : R. Schmalensee and R.D. Willig (eds.), Handbook of Industrial Organization, Elsevier Science Publishing : Amsterdam. Williamson, O.E. [1991], Comparative economic organization : the analysis of discrete structural alternatives, Administrative Science Quartely, 36, 269-296

18