Determinants of the decision to outsource information systems in the Netherlands.

(1)

Determinants of the decision to

outsource information systems in

the Netherlands.

University : Rijksuniversiteit Groningen

Faculty : Economics

Specialization : Economics & Management

Author : Keur, Ron

Student number : 1257382 First Supervisor : Dr. G. Péli Second Supervisor : Dr. H.F. Lanting

(2)

IS Outsourcing Determinants. R. Keur (2006)

Abstract

It is the goal of this study to examine Dutch organizations’ motivations with respect to their IS outsourcing decision. This study focuses on Dutch based firms because an increasing trend is discernible in organizations in the Netherlands in the amount of outsourced information services. Furthermore, literature lacks empirical studies conducted on Dutch based firms, therefore, this paper attempts to fills a gap in the literature.

This study presents an empirical analysis of the motivations of organizations to outsource their information systems. With the results from this analysis, researchers can reflect on the question whether the decision to outsource was based on correct assumptions and expectations from managers. When expectations from managers deviate from realistic scenarios, expectations have to be altered and other decisions may be the outcome. When managers were too positive or had wrong reasons for outsourcing, it might not be optimal to outsource at all and it is better to keep the IS function in-house.

Major conclusions of this study, are that Dutch based firms appear to outsource their information systems because they want to focus on their core activities. Furthermore, cash injection is also a supported motivation for the outsourcing decision. Motivations as potential cost reductions and general low level of business performance are not supported in this study.

1. Introduction

1.1 General background

(4)

Dutch based companies follow this trend as well. In 2004, turnover generated by outsourcing contracts in the Netherlands was good for a fifth place in Europe (See figure 1). The services that were outsourced most were the management of IT infrastructure, followed by application management and desktop management (Computable.nl, 1-7-2005).

What is remarkable is that European contracts tend to be smaller (mostly up to 20 million Euros) than North American contracts (mostly over 20 million Euros). Furthermore, contracts in the Netherlands are more short term oriented (less than 5 years) than they are in other European countries, where long term contracts are gaining popularity (Computable.nl, 1-7-2005).

The question that rises, is why organizations are making increasing use of external IS suppliers? Markets are becoming more and more competitive and managers seek ways in which to improve their business processes and efficiencies. Outsourcing certain parts of the organizations can increase business performance. The outside IS services management is even indicated by Clark (1992) as one of the six strategic management issues confronting organizations. Managers face a trade-off though, when deciding whether to outsource their IS department. On the one hand, several arguments in favor of outsourcing exist. For example, external suppliers are believed to be able to provide a higher quality service at lower cost. Furthermore, client organizations are expected to be better able to focus on their core activities once their IS function is outsourced. The cash injection related to the selling of the IS hardware can also be an argument in favor of outsourcing.

On the other hand, there are also some major drawbacks to outsourcing the IS function. First, an organization might become excessively dependent of the IS supplier. Second, critical house skills and competences are lost and it is therefore difficult, once outsourced, to

in-Figure 1: Turnover in IS outsourcing deals, Europe, 2004 (in billions of euros)

(5)

source the IS function once again. Finally, since the vendor has to make a profit margin on its service and this is often charged to the client, a cost advantage might not even exist.

A recent survey by Gartner group among organizations that outsourced their IS, indicates that 50% of the managers were not satisfied with the results from outsourcing (Computable.nl, 4-4-2003). In figure 2, the results of a study by Giarte Media Group can be found (Computable.nl 25-10-2005). They found managers’ satisfaction levels on specific outsourced processes. Again, managers seem to be disappointed with the obtained results from the contract. These studies indicate that there is a discrepancy between the expectancies managers have around outsourcing information systems and the actual results they see. It is important for managers to align their expectancies with realistic scenarios. It is therefore interesting to see what the reasons and motivations behind an outsourcing decision were, and whether they are congruent with realistic scenarios.

It is the goal of this study to examine Dutch organizations’ motivations with respect to their IS outsourcing decision. This study focuses on Dutch based firms because an increasing trend is discernible in organizations in the Netherlands in the amount of outsourced information services. Furthermore, literature lacks empirical studies conducted on Dutch based firms, therefore, this paper attempts to fills a gap in the literature.

This study presents an empirical analysis of the motivations of organizations to outsource their information systems. With the results from this analysis, researchers can reflect on the question whether the decision to outsource was based on correct assumptions and expectations from managers. When expectations from managers deviate from realistic scenarios, expectations have to be altered and other decisions may be the outcome. When

Figure 2:

Satisfaction levels with outsourcing processes

89% 70% 59% 52% 52% 51% 50% 48% 47% 75% 62% 65% 59% 40% 56% 58% 0% 20% 40% 60% 80% 100% BPO External Networks Hosting business applications Functional application & data

management Technical application & data

management Helpdesk Application development &

integration Office automation

(6)

managers were too positive or had wrong reasons for outsourcing, it might not be optimal to outsource at all and it is better to keep the IS function in-house.

1.2 Theoretical background

This section is organized as follows. First, the definition of IS outsourcing which is applied in this study will be highlighted briefly. Furthermore, a brief overview of some relevant literature concerning the motivations of the decision to outsource information systems is presented. The first part discusses some theoretical points of view where agency theory and transaction cost theory form the centre of attention. The second part discusses some empirical results from studies conducted in other countries. Finally, I will elaborate on the benchmark paper of this study, which is conducted in the United States by Smith et al. in 1998.

In this study, a widely accepted definition for the term IS outsourcing is used. It is important for the analyses presented here that a clear definition of IS outsourcing is in use. Cases which are to be selected, should comply fully with this definition in order to make a clear distinction between organizations that did, and organizations that did not outsource their IS. The definition used here is derived form Gonzalez et al. (2005) and is implicitly adopted by many other authors on the topic1. The definition used is the following:

“IS outsourcing means that the physical and/or human resources related to an organization’s information technologies (IT) are going to be provided and/or managed by an external specialized supplier. The situation can be temporary or indefinite and can affect the client firm’s whole IS, or only part of it” (Gonzalez et al., 2005).

Theoretical perspectives

Ngwenyama et al. (1999) analyzes the decision to outsource using a transaction cost approach. In this study, potential costs specifically associated with IS outsourcing are analyzed. Transaction cost theory should determine the most efficient mode of organization: performing a function in house or let the market provide it to you (Williamson, 1976). The

1

(7)

authors state that the primary objective of managers making the IS outsourcing decision is to minimize total cost (service and transaction costs) and thereby maximizing total firm value. In this manner, advantages of IS outsourcing are described as a minimizing cost function. Ngwenyama et al. (1999) has broken down specific transaction costs associated with the outsourcing of the information function in four categories. The first is the cost of the information processing service itself. These costs can be reasonably estimated from the market. The second type of cost is set-up or contracting costs. These costs include search costs to find a suitable vendor, negotiation costs, legal fees and other costs related with the institutionalization of the relationship. These costs can be quite high because information asymmetry and uncertainty surrounding information systems is usually large. The third type of costs is incurred by monitoring and coordinating the relationships and contracts. Again, since information asymmetry is high, and it is thus difficult to monitor the agents’ actions, these costs can be high as well. The final type of costs is switching costs which are incurred when a client firm tries to switch vendor. Since a vendor usually has much inside information on the business of a client organization and a relationship already exists, these costs can often also be assessed as high.

Since transaction costs are potentially high with respect to letting the market perform the IS function, managers have to consider carefully whether transaction costs are minimal when the IS function is outsourced. In other words, if the advantages of retaining the function in house do not outweigh the advantages associated with outsourcing. This study tries to elaborate on the advantages associated with outsourcing the IS function in the Netherlands. When the associated advantages of IS outsourcing are known, these can be compared to potential disadvantages thereby attempting to minimize transaction costs.

Cheon et al. (1995) define a conceptual and a contingency based model for studying the motivations of outsourcing information systems. The model incorporates four theories which together should define the desirability of outsourcing. These in turn can be seen as potential motivations to outsource the IS function.

(8)

Another light in which the decision to outsource can be seen, is the resource dependence theory. This theory states that organizations should adapt their strategy in order to gain access to scarce resources in the environment of the firm. Outsourcing IS is a strategic decision, and organizations should outsource because it provides them access to these resources.

Two final theories which are incorporated in the model are transaction cost theory (already discussed above) and agency cost theory. When agency cost theory is applied to IS outsourcing, the goal of the manager is to minimize the agency cost of the chosen relationship (internal or external organization of the activity). Agency costs in turn depend on uncertainty, risk aversion, programmability, measurability and length of the contract. Agency costs are typically high in IS outsourcing deals and managers have to take these costs into account. The insights given by Ngwenyama et al. (1999) and Cheon et al. (1995) with respect to IS outsourcing motivations, can provide a general direction for this empirical study. Results uncovered by these theories can also be found in other empirical works described here.

Empirical perspectives

The studies mentioned in this section, can again provide a general direction for the research question in this study. All the following papers performed an empirical analysis on IS outsourcing motivations. Studies varied in countries in which they were performed and in methods used to obtain the results. The results these studies obtained were all in the form of certain motivations of organizations to outsource their information systems.

Cross (1995) performed a case study on the IS outsourcing process at British Petroleum. Cross found that BP outsourced all its information technology operations in an effort to cut costs, gain more flexible and higher-quality IT resources, and focus the IT department on activities that directly improved the overall business (Cross, 1995).

Currie (1998) discusses two British case studies in which the process of outsourcing is evaluated. The first organization was ICI (a chemical company). Main motivation at this company to outsource their IS was to cut cost and rationalize their business (i.e. return to their core activities). The other company was Wessex Water (a water company). The key driver of the decision here, was that Wessex Water wanted to focus on their core competencies.

(9)

at serving customers. Access to new technologies and more qualified staff was the key driver in this case.

Fowler and Jeffs (1998) performed a case study in the UK as well. Most important result in this study was that though rational motivations were given by managers, irrational and political motivations when making the decision seems to have been given more weight. Though political motivations have been discussed in literature (see for example Lacity et al., 1994), not much empirical studies have been conducted on this topic.

Gonzalez et al. (2005), conducted a survey with a final sample size of 357 firms. The survey asked managers in the organizations the major motivations they had for outsourcing their information systems. The results of this survey were interesting. The most important quoted motivation was that the organization wanted to focus on IS strategic issues and not on commodity issues. This motivation was followed by a desire to improve IS department flexibility and to improve IS quality. It is remarkable that cost savings or the correction of technological deficiencies were far less important motivations for managers in this study. When examining the available literature, two major motivational factors can be distilled; economic factors and strategic factors. Other factors can influence the decision as well, however, these two factors appear to capture the main drivers of IS outsourcing decisions.

Economic Factors

Cost reduction is the most quoted rationale for IS outsourcing (Clark et al., 1995). Organizations outsource their information systems because they believe that an outside vendor can achieve economies of scale and scope with respect to the IS function. It is a common assumption that IS service providers are in a better position to exploit these advantages since it can pool hardware and staff for many service receivers, has better access to low cost labor and has more expertise in controlling typical IS costs (Grover et al., 1994). The actual realized cost reductions are a point of discussion though. This is because vendors often use the same staff and hardware as before, and needs to make a profit on top of that (Lacitiy & Hirschheim, 1992). Hidden costs in the contract can also be an obstacle in realizing cost savings (Lacity & Hirschheim, 1994).

(10)

to costs more directly. Furthermore, every IT related decision which can lead to extra costs has to be submitted through a formal cost control system. It remains the question though, if an increased cost control cannot be obtained by installing more or better cost control systems while still retaining the IT function in-house.

Another economic factor which can affect the decision to outsource, is an organizations’ need for liquidity. Often, the outsourcing contract involves an initial payment from the vendor for tangible and intangible IT assets from the client (McFarlan & Nolan, 1995). In this manner, non liquid assets, such as a mainframe computer, can be turned into liquid assets more rapidly. Organizations hereby transform a long term fixed capital outlay into periodic payments to the vendor. This increases a firms’ cash flow, frees up cash for other, potentially more lucrative, investments and gives the company the possibility to repay outstanding debts. A disadvantage in transforming non liquid assets into liquid assets in such a manner, is that vendors usually require long term contracts to recoup on their investment. Since fees are often determined up front of the agreement but last for the entire length of the agreement, client organizations do not share in the improvements in price/performance ratios which are bound to exist after a number of years (Lacity & Hirschheim, 1994).

Strategic Factors

The main strategic motive for organizations to outsource their IS, is to simplify the management agenda and focus on the core activities (Grover et al., 1994, Lacity & Hirschheim, 1994). Zhang & Cao (2002) expresses this line of argument as follows:

“To succeed, corporations need streamlined and globally integrated operations. Above all, they need an absolute focus on their core business. That is when it may become important to outsource certain “non-core” requirements.”

(11)

Gupta & Gupta (1992) state that information is a strategic resource for organizations, no matter what line of business they are in. Managers should therefore carefully consider the option to outsource a potential strategically important part of the organization to an external supplier. It is important for managers to distinguish between commodity like type of IS provision and IS provision with a true strategic nature and they should retain the business critical IS functions in-house.

Another strategic motivation, is a general low level of performance of the organization. Strassmann (1995) remarks the following:

“Outsourcing is in reality only one aspect of a currently popular downsizing trend among troubled corporations”.

In other words, firms with low profitability and weak performance see outsourcing their IS function as a way to cut employment and get rid of unneeded in-house capacity. Furthermore, under conditions of poor business performance, organizations often seek to streamline operations, including selling of or redeploying assets (Harrigan, 1980). From this perspective, outsourcing is merely a means in an effort to turn the company around into a profitable one.

Smith et al. (1998) performed an empirical study in the United States based on the factors aforementioned. More specific, motivations as cost reductions, focus on core competence, cash needs and a general low level of performance were examined on the base of financial characteristics of organizations that outsourced their IS. This study introduced four testable propositions with respect to these key drivers of the IS outsourcing decision.

Their study incorporated 29 US based firms who outsourced their information systems in the period 1985 – 1995. These firms were analyzed on the base of a number of financial indicators that would reveal the presence of certain motivations. The most important results of the study were that support was found for the proposition that firms who outsourced IS were more cost conscious than other firm and that firms outsourced their IS for cash needs. There was however, no support found for the propositions that sample firms were focusing on their core competence or that sample firms had lower profitability.

(12)

assumed that companies in the Netherlands do not significantly differ from US based companies. It can be interesting to see if motivations found in the US based firms are different from Dutch based firms. Discrepancies in these results can be the consequence of differences in corporate culture, business environment or governmental policy and might provide interesting topics for future research.

1.3 Research objectives

The goal of this study is to examine the motivations managers have when the outsource their information systems. More specifically, this study seeks to find empirical support for a number of key drivers of IS outsourcing in organizations based in the Netherlands. Support is examined by studying financial characteristics of firms who have entered large scale IS outsourcing arrangements. Similar research is done by Smith et al. (1998) in the US. This study will build on their methods and justifications. In this manner, comparability between the two studies (and thus the two countries) is possible.

This study aims at organizations based in the Netherlands. This is because it can be interesting to see whether these exist differences in motivations to outsource between countries. Empirical work is conducted on a large scale in a number of countries. Smith et al. (1998) examined motivations in US based firms, Gonzalez et al. (2005) studied a large number of Spanish firms and Fowler & Jeffs (1998) performed a study in the UK. All of these studies typically show some differences in their results. It is interesting to see how Dutch firms fit in this picture and why it does so.

Smith et al. (1998) tests four propositions with respect to IS outsourcing motivations. Similar propositions are incorporated in this study as well. The four propositions are based on the key drivers of IS outsourcing discussed earlier.

(13)

often highlight other motivations as key drivers. Currie (1998) did find cost cuttings as one of the (though not a main motivation) motivations for IS outsourcing at ICI. Other literature supporting the cost reduction motives are often more theoretical in nature.

Based on economic rationale and on the results obtained by Smith et al. though, it can be expected that organization who outsourced their IS, had a tendency to be more focused on cost reductions (they had a greater need to cut costs) than other firms in their industry. The following proposition to test in the Netherlands is therefore introduced.

Proposition 1: Organizations that outsourced (most of) their IS had a greater need to cut costs than other firms in their industry.

Another key driver of IS outsourcing was an organizations’ need for cash. Smith et al. found empirical support for the idea that companies who outsourced IS were in greater need for cash than other firms. This proposition is also supported by some theoretical work from Alner (2001) and Clark (1995). To test the validity of this key driver in the Netherlands, the following proposition is put forward.

Proposition 2: Organizations that outsourced (most of) their IS had a greater need for cash generation than other firms in their industry.

(14)

Proposition 3: Organizations that outsourced (most of) their IS had a tendency to focus more on their core activities than other firms in their industry.

The final proposition, deals with an organizations’ need for improvements in profitability and performance. Strassmann (1995) believes that outsourcing is way for low performing organizations to increase their performance. Smith et al. (1998) did not find support for the assertion that companies who outsourced their IS were systematic underperformers in their industry. Because their is some economic rationale, it is expected that organizations outsource their IS because they have a greater need to increase profitability or organizational efficiencies in order to improve organizational performance. The following proposition is incorporated to test this assertion in the Netherlands.

Proposition 4: Organizations that outsource (most of) their IS performed worse than other firms in their industry.

The number of propositions which can be stated are not exhaustive. However, the presented propositions are a reflection of the theoretical major determinants of IS outsourcing decisions. Furthermore, these propositions can be tested in a reasonable manner by analyzing publicly available financial data. Finally, since the propositions resemble propositions from earlier studies in other countries, Dutch firms can be compared with other firms.

2. Methodology

This study seeks to find empirical support for some theoretically expected key drivers of IS outsourcing of companies based in the Netherlands. Support is sought by comparing firms who have entered into large scale IS outsourcing contract with their respective industry counterparts. More specifically, firms are compared on some financial characteristics.

(15)

In this section the methodology of the research process will be described. First, the way in which cases were selected is clarified. Thereafter the methodology of quantitative data collection is explained. Finally, the way in which the collected data will be analyzed will be exemplified.

2.1 Sample Case Selection

Sample Cases

For the study, a case sample selection had to be made of companies who outsourced their information systems. Cases were selected using the following procedure.

A newspaper databank (Lexis Nexis) and an ICT news website (www.computable.nl) were used to search for announcements of companies who were going to outsource their IS. Only large scale outsourcing contracts were selected. These cases were selected by searching on key words such as large scale outsourcing, complete outsourcing, network management etc. It was important in the selection process to exclude cases where IS was outsourced only partially. It is assumed that the most drastic cases of outsourcing provide the best evidence in support of the propositions.

Furthermore, the sample cases had to be headquartered (or completely based) in the Netherlands to account for macro economic effects. Dutch based cases were also a criterion in order to be able to compare with other counties and prior research.

A final criterion in the case selection was the availability of financial data. Public organizations and privately held companies were thus excluded from the sample.

The result of the selection process was a sample of 19 firms which all outsourced their IS in the period 2000 – 2005. For these cases, the year in which the decision to outsource was announced was labeled year 0.

(16)

Industry Control Groups

Four each firm a control group was also established to correct for industry effects. The control group sample was selected on the base of the same 3 digit BIK code as the sample firms’ main activity. Furthermore, the control group consisted only of firms of which financial data was readily available. When the control group became too large for analysis, the largest firms (based on their numbers of employees) were selected as peer group. Peer groups varied in size from only 4 peer group firms to 35 peer group firms.

2.2 Operationalization

This study seeks to find empirical support for theoretical key IS drivers in the Netherlands. Many studies have approached this problem by adopting a case study approach. In these instances interviews and questionnaires were collected in order to give a complete picture of a firms’ rationale for outsourcing. In this study however, empirical support is examined by studying publicly available financial data and financial reports. The latter approach has several advantages. First, the data with respect to sample case firms and industry averages is easier to obtain. Second, when drivers are conceptualized properly, the results provide a more objective picture than single in depth case studies. A final advantage is the comparability of financial data over different companies, sectors or countries.

A disadvantage of studying publicly available data is that the research is heavily dependent on what data is available. Because of this, financial metrics which may only have an indirect relation with the real motivation to outsource can be selected. Furthermore, the data which is studied covers an entire organization. Patterns discovered in this data might be the result of other organizational phenomena than the assumed relationship.

For these reasons, it is important to derive a set of well contemplated financial metrics which describe the true motivation for IS outsourcing best as possible. Since the metrics can only measure the true motivations indirect, multiple metrics per driver have to be selected to validate the results. A summary of the financial metrics which shall be used can be found in table 1.

(17)

outsource. Metrics are also analyzed for trends. Trends in certain metrics prior to the outsourcing decision might also be an indication of a firms’ motivation to outsource.

Table 1: Summary of financial metrics

Driver Measurement Definition

Expectations for sample firms Cost Reduction Operating expense

ratio Operating expense / Sales Higher

Debt ratio Total liabilities / total assets Higher

Financial leverage Total liabilities / total SE Higher

Growth rate S(x) - S(x-1) / S(x-1)* Lower

Liquidity

needs

Liquidity Liquid assets / total assets Lower

Debt ratio Total liabilities / total assets Higher

Financial leverage Total liabilities / total SE** Higher

Focus on core

competence

Sales per employee

Sales / number of

employees Lower

Profit per employee

Profit / number of

employees Lower

ROA EBIT / total assets Lower

Improve

performance

Efficiency Sales / average total assets Lower

Profit margin EBIT / sales Lower

ROE EBIT / total SE** Lower

ROA EBIT / total assets Lower

* Sales

(18)

Cost reduction

Proposition 1 states that firms who enter an IS outsourcing contract do so because they seek to reduce costs. From this starting point, I assume that these firms are more cost conscious than other firms in the same industry. This implicates that the sample firms, because they are more cost conscious, perform better on certain financial metrics. This proposition can be tested by comparing the following measures.

First, a cost efficiency metric is introduced by which the proposition can be tested directly. Operating expense normalized by a firm’s level of sales is a good proxy of measuring a firms’ cost of operations. It is expected that firms outsource their IS in order to improve on their cost position. It is therefore logical to expect that at the moment before outsourcing their relative cost structure is unfavorable compared to their industry counterparts. This metric is thus expected to be lower for the sample case firms.

Other, indirect support for this proposition, can be drawn from other measures such as the amount of debt and the level of growth. In the finance literature, debt is often seen as a means of cost control. The main argument used here, is that managers with an excess of future free cash flows will invest this cash in negative-NPV projects. The creation of debt limits the managers’ control over the future cash flows by obliging them to repay their debts (otherwise debt-holders can take the firm into bankruptcy) (Jensen, 1986, Harris & Raviv, 1999).

Since managers now face a greater risk when investing in negative NPV project, managers are more likely to run a more efficient operation and consider more carefully before spending money (Brealey et. Al. 2001).

The level of debt can be measured by the debt ratio (total liabilities as a percentage of total assets) and by the financial leverage (total liabilities as a percentage of total shareholders equity) of a firm. Higher debt burdens and high levels of financial leverage indicate a more cost conscious firm.

(19)

increase profits. It is thus expected that the sample firms have lower or more declining growth rates compared to the industry average.

Liquidity needs

The second proposition proclaims that organizations outsource their IS for a cash infusion (which is often part of the deal). This hypothesis can be tested by examining the degree of how much organizations are in need of cash. This need of cash can be measured using a number of metrics.

First, a companies’ level of liquidity may be an indicator of their need for cash. Companies with low levels of liquidity, might need a quick infusion of cash in order to provide for short term operating needs. The liquidity of a firm can be measured through the ratio of liquid assets of a firm to its total assets. It is expected that the sample firms are in need of cash and thus exhibit relative low levels of liquidity.

The amount of debt a company carries, can also be an indicator of a need for cash. Organizations may need cash in order to reduce large debt burdens. Debt levels of a firm can be measured in several ways. In this study the level of debt is examined by the debt ratio (total liabilities as a percentage of total assets) and the financial leverage of a firm (total liabilities as a percentage of total shareholders equity). Sample firms are thus expected to carry higher levels of debt and financial leverage than their industry counterparts.

Focus on core activities

Proposition 3 states that organizations who outsource their IS, are in an organization wide effort to focus on their core activities by outsourcing non-core activities. It is difficult to measure the degree of a companies’ focus by interpreting publicly available financial data. There are however indirect measures available by which a companies’ focus can be determined.

(20)

organization. In this study this measurement is used to determine to what degree organizations focus on their core competences.

Four financial measurements were tested on manufacturing organizations to test their degree of virtuality. The first two measurements that were used relied on the underlying assumption that virtual organizations have relatively little direct employees (Presley et al., 2001). It is expected that firms that outsourced many (non-core) activities employ less people because most of them work for another employer. From this assumption it follows that virtual organizations should exhibit higher ratios of sales per employee and profit per employee, relative to more traditional operating organizations. The sample case firms probably exhibit lower ratios in the years prior to outsourcing, because they are expected to be in an effort to streamline their organization.

A third metric discussed, assumes that virtual organizations have relative low levels of stock and work in progress. This metric however, is mainly suited for manufacturing organizations. Since this study is also aimed at more service oriented organizations, this metric will not suffice.

A final metric discussed, is a firms’ Return On Capital Employed (ROCE). This measure assumes that virtual organizations have relative little capital employed for similar levels of profits. Since not all data on the capital employed by organizations is available, a similar measure (the ROA) shall be used to define a companies’ ability to obtain similar levels of profits for fewer investments in assets. It is hypothesized that virtual organizations have to invest less in assets than traditional organizations, and thus obtain higher levels of ROA. Again, since the sample case firms are trying to become more virtual, ROA levels are expected to be lower for the sample case firms in the years prior to outsourcing.

Improve performance

The fourth proposition states that companies outsource their IS with the rationale of improving organizational performance. Organizational performance in turn, can be captured by concepts of efficiency and profitability.

(21)

expressed as the profit margin, the return on equity and the return on assets, and is again expected to be lower for the sample firms as well.

The measures of efficiency and profitability can be interpreted as general measures of organizational performance. It is therefore expected that the sample firms exhibit lower organizational performance in comparison with other firms in the industry.

2.3 Data collection

For each firm in the sample, financial data had to be gathered. Balance sheets and income statements 3 year prior to the outsourcing decision had to be available. This financial data was gathered from the REACH database in which data from 1.7 million organizations registered at the chamber of commerce in the Netherlands is available.

For each sample firm a table is generated which contains the financial ratios for the 3 years prior to the outsourcing decision. For each sample firm, a control group table which contains the medians from the control group data was also created. The median was used because this number is less influenced by outlying data points. For each firm and each metric, the difference was calculated between the sample firms’ variable and the industry median. This resulted in a list of 76 (4 years times 19 firms) positive and negative differences for each metric. It is this list which will be analyzed to detect differences between sample firms and industry averages. Changes in metrics from years -3 to -1 and from years -2 to -1 are also measured to reveal certain trends in the metrics.

2.4 Data Analysis

(22)

signed rank test, tests if the number of firms that perform better than the industry median, is significantly different from the number of firms that perform worse than the industry median. Furthermore, a trend in the metrics for each firm is also analyzed. Differences between years -3 to -1 and between -2 to -1 will be analyzed using the same Wilcoxon signed rank test. The Wilcoxon signed ranked test is performed as follows2. First, the Wilcoxon signed rank test defined the differences of each specific subject:

Di = Xi – Yi for i = 1, ...,n. (Items for which Di = 0 are excluded from the test).

The next step is to rank the differences of each specific subject, subjects with the smallest deviation from 0 are assigned rank 1. The following null-hypothesis is that the sum of the ranks of subjects with a positive deviation (T+) does not differ from the sum of the ranks of negative deviations (T-), or, in other words that the two groups do not differ from each other. All tests are performed in Excel.

2.5 Limitations

The chosen method of research has several drawbacks with respect to the selected financial measurements. The most important disadvantage of studying publicly available financial data, is that the evidence obtained can only provide indirect support for the propositions. This means that a firms motivation to outsource their IS may not be revealed by their financial characteristics. To minimize this effect, a number of precautions were taken. First, the metrics chosen were based on well established theories in finance and strategic management. Second, multiple financial metrics per proposition were used and evidence from multiple metrics would support the propositions even more.

Another problem which may arise, is that external conditions may exist that influence the metrics but are not related to the motivation for outsourcing. To correct for this problem, metrics are adjusted to account for industry effects. When a firms’ metric is subtracted from its industry median, the industry effect is (partially) removed.

2

(23)

3. Analysis and results

This section presents and analyses the empirical results found on the four propositions. First, some points of interest regarding the data based used are raised. Then the empirical results are presented.

Because the financial metrics presented in this study can only provide indirect support of the drawn propositions, evidence is sought from multiple metrics on each proposition. I believe that support for a hypothesis on a certain financial metric exist, when the signs of every industry adjusted figure is the same in all analyzed years and when at least two of them are significant at a 10% level of significance. Furthermore, support for the propositions is believed to exist when at least two hypothesis on the chosen financial metrics are supported.

3.1 Sample firms database

The sample case database started with 19 firms in the selection. One of these firms, Gasunie, has been split up two years prior to the date their outsourcing decision was made. Since only one of the split up firms chose to outsource their information systems the obtained data was distorted: prior to split up only financial data for the whole of Gasunie was available. This made a longitudinal analysis for Gasunie difficult. Finally, this difficulty resulted in outlying data points for almost every year for Gasunie, rendering the obtained data useless. It has therefore been chosen to leave Gasunie out of the sample.

Furthermore, ING (one of the sample case firms) chose to outsource their information systems in 2006. For this year, no financial data was available at the time of the study. Subsequently, financial data of ING has been gathered, however, year 0 data for ING is lacking.

(24)

3.2 Empirical results

Proposition 1: Cost reductions

The table below (table 2) presents the results for the financial metrics for the first proposition. The first tree columns represents the differences between the sample firms’ value for a certain metric and the industry median value for the same metric in their corresponding years. For example, the median operating expense ratio for year 0 for the sample case firms was 1,2% higher than its median industry counterparts. The final two columns present the trends from years -3 to -1 and years -2 to -1. These values are obtained by subtracting the year -3, respectively year -2, industry adjusted value from the -1 year industry adjusted value. Values again are shown as percentages. The levels of significance are based on a two-tailed Wilcoxon signed rank test. A firms’ incentive to induce cost reductions is operationalized by 4 measures: Operating expense, debt levels, financial leverage and growth rates. As can be seen from table 1, the evidence in support of proposition 1 is not sufficient.

The operating expense ratio is indeed higher for the sample firms for every year, as was hypothesized. The differences are however significant in only one year. A negative trend in this metric for the sample case firms can also be perceived, however again not significant. It therefore does not appear that sample case firms exhibit higher levels of operating expense for the same levels of sales relative to their industry counterparts.

Table 2: Cost reduction metrics

Year -3 -2 -1 0 -3 to -1 -2 to -1

(25)

The debt ratio of the sample firms does not seem to differ from it’s industry median at all. Signs for the differences differ across years and none is statistically significant. While a positive sign is expected for this metric, three of the four year are negatively signed. Signs for the trend for this metric show that sample firms are having declining debt ratios prior to outsourcing, though again not statistically significant. It can therefore be concluded that firms that outsource their information systems do not exhibit higher debt levels than their industry counterparts.

There seems to exist only minor support for the hypothesized difference in financial leverage from sample firms to industry medians. Signs for the metric are, as expected, positive for the sample firms. However, the result is only significant on a 10% level in one year. No significant trend can be discerned in this metric either. Again it can be concluded that no difference exist between sample firms and other firms with respect to their levels of financial leverage.

The final metric with respect to proposition 1, the sample firms’ growth rate, does not seem to differ between sample case firms and industry medians. Signs differ across years (three of four years negative) and only one year is statistically significant at a 10% level. Yet again, no significant trend can be discerned for this metric. It thus appears that sample case firms do not exhibit lower growth rates compared to their industry counterparts as hypothesized.

All in all, support for the first hypothesis weak. Results are inconsistent and only significant at a low level. Sample case firms therefore do not seem to have a greater need to cut costs than their industry counterparts.

Proposition 2: Liquidity needs

A firms’ need for liquidity is earlier defined with 3 financial metrics: a firms’ liquidity, its debt ratio and the level of financial leverage. The results for the sample case firms liquidity needs compared with their industry counterparts can be found in table 3.

(26)

though. For this metric it can be concluded that sample case firms do have a relative worse liquidity position compared with their industry counterparts, which was hypothesized.

The hypothesis that sample case firms carry higher debt ratios, is not supported by the empirical results (as with the first proposition). Signs differ across years and none is significant. What’s more sample case firms do not appear more leveraged than their industry counterparts (as was the result with the first proposition).

Table 3: Liquidity needs metrics

Year -3 -2 -1 0 -3 to -1 -2 to -1 Liquidity -0,055 * -0,059 -0,213 * -0,299 ** -0,056 -0,243 Debt Ratio -0,013 0,027 -0,022 -0,012 -0,370 -0,236 Financial Leverage 0,121 0,331 0,115 0,443 * -0,321 -0,300 *** Significant at 1% level ** Significant at 5% level * Significant at 10% level

All in all, there is little empirical support found for the second proposition. Though sample case firms exhibit lower levels of liquidity, debt ratios and financial leverage do not differ from their industry counterparts. Sample case firms thus not appear in greater need of cash than other firms.

Proposition 3: Focus on core activities

A firms’ measure of virtuality is defined here as 3 financial metrics: sales per employee, profit per employee and a firms’ return on assets. Table 4 presents the empirical results with respect to the third proposition.

The third proposition finds some very appealing empirical support. All the metrics are significant on at least a 5% level two years prior to the outsourcing decision, and two of them are significant at the 1% level. All the signs the metrics exhibit, are as what was hypothesized.

(27)

and in year 0 at a 1% level. There is some slight support for a negative trend in this metric as well. These results implicates that sample case firms need more employees to generate the same amount of sales.

The same holds for the second metric, profit per employee, which is statistically significant in 2 years. In year -1 at a 1% level and in year 0 at a 5% level. Again, some slight support of a negative trend in this metric is found. I conclude that sample case are less profitable per employee relative to their industry counterparts.

Table 4: Focus on core activities metrics

Year -3 -2 -1 0 -3 to -1 -2 to -1

Sales per employee -0,033 -0,241 ** -0,197 ** -0,285 *** -0,015 ** -0,043 Profit per employee -0,091 -0,242 -0,451 *** -0,336 ** -0,305 * -0,016 * ROA -0,085 -0,030 -0,402 ** -0,218 ** 0,318 0,001

*** Significant at 1% level ** Significant at 5% level * Significant at 10% level

The final metric, return on assets, exhibits similar results. The metric is negative in all years an statistically significant at a 5% level in the two years prior to outsourcing. No statistically significant trend can be discerned for this metric though. However, it can be concluded that sample case firms need more assets to generate the same amount of profits as their industry counterparts.

(28)

Proposition 4: Performance improvement

The final proposition with respect to firms’ performance levels is subdivided in four financial metrics: efficiency, profit margin, return on equity and return on assets. The results for these metrics are presented in table 5.

The hypothesized difference with respect to levels of efficiency does not seem to hold in this study. Signs of the results differ across years and none is significant. There also does not appear to be any trend. Sample case firms, so it appears, do not to operate less efficient than their industry counterparts.

Table 5: Performance improvement metrics

Year -3 -2 -1 0 -3 to -1 -2 to -1 Efficiency -0,035 0,069 0,164 -0,084 0,074 -0,025 Profit margin -0,065 -0,021 -0,277 ** -0,157 -0,352 -0,171 ** ROE -0,049 0,139 -0,286 0,023 -0,250 0,031 ** ROA -0,085 -0,030 -0,402 ** -0,218 ** 0,318 0,001 *** Significant at 1% level ** Significant at 5% level * Significant at 10% level

Profit margins are consistently lower for sample case firms for every year, and in one case significant on a 5% level. A negative sign in the trend for this metric can also be discerned and is significant on a 5% level in one case as well. As a result, there does seems to be some empirical support for the hypothesized difference between sample case firms profit margins relative to their industry counterparts.

The results on the third metric, return on equity, again are not consistent over the years, and statistically not significant. Sample case firms therefore appear to be no different from their industry counterparts when it comes to their return on equity.

(29)

Support for the fourth proposition, that firms who outsource their information systems generally under-perform the market, is weak. Two of the metrics can be interpreted as slightly significant and show some support in favor of the proposition. However, the other two metrics do not show any support for the hypothesized relation.

4. Discussion and Conclusion

This study sought to find empirical support from publicly available financial data for motivations for firms to outsource their information systems that emerged from the literature. Motivations that arise from the literature are translated into 4 propositions which in turn were subdivided in a number of financial measurements. This section will evaluate the propositions in the light of some previous relevant literature and the empirical results from this study. A comparison with a benchmark study based in the United States will also be made. Furthermore, some implications for managers and some limitations of the results will be presented. The chapter ends with a conclusion.

4.1 Discussion

Proposition 1:

Cost cutting is the single most quoted reason in the literature that exists as motivation for companies to outsource their information systems. It is generally believed that outside vendors can provide the same levels of service and quality for lower costs. Theory stresses that economies of scale and access to cheap labor are the sources of this cost advantage. The cost cutting rationale is also highlighted by a number of case studies.

(30)

avoid situations of financial distress. Outsourcing their information systems may be the cost cutting measure managers seek. The final examined motive for cost cutting, are low growth rates. When organizations suffer low growth rates in sales, the only way to increase their profits is to cut costs. In all of these situations cost cutting can be a motivation for the manager to turn to outsourcing their IS.

When looking at the results of this study, it appears that cost cutting is not a major motivation for managers to turn to outsourcing their IS (see table 2). Results show that companies who have outsourced their IS do tend to have higher level of operating expense relative to their industry counterparts, though statistically not very significant. Furthermore, sample case firms were more leveraged, though again, results are not significant. The other two measures which were studied, debt levels and growth rates, did not show any consistent support. Though the first proposition cannot be rejected by the evidence presented, it also cannot be confirmed either.

When examining previous case studies and literature reviews, evidence for the cost cutting motivation is overwhelming. This raises the question why these results aren’t supported by the empirical study conducted here. A number of reasons for this inconsistency can be taken into account.

It is possible that managers have cost cutting as a high priority when they outsource their IS. However it is not necessary that the results of the metrics shown here present a managers’ true motivation. It might very well be that managers think they have to cut cost while their balance sheet or income statement does not offer a clear cut reason for this line of action. When for example a company strives to be a cost leader, cost cutting are always sought, regardless of the current financial state of the organization. Fierce competition and cost based reward systems can also be motivations for managers to cut costs and thus outsource their IS. These underlying motivations are more political in nature and can be difficult to measure while analyzing publicly available financial data.

(31)

contributors of the poor results found. Can lower levels of debt for these firms be the result of other factors than a firms’ goal to limit a managers’ control over future cash flows? PCM for example re-financed its organization with greater levels of debt one year after they outsourced their IS and the same holds for KPN. Philips has low debt levels as one of its main objectives in order to save on interest costs. So, not all organizations seem to use debt financing in order to control cash flows. Furthermore, when organization witness a sudden increase in sales with equivalent costs, the operating expense ratio will decrease. This happened with organizations such as Heijmans, DSM and ING. These facts distorts the data in such a way that results obtained were rendered not significant. For future research, more fine tuned financial metrics have to be used or sample case firms have to be selected on the goal of their debt financing. The hypothesis that sample case firms should exhibit lower growth rates, has also some shortcomings. First, if an organizations’ sales decline in one year, this does not necessarily imply that managers are panicking and start outsourcing parts of the organizations. There may be business cycles at work which can be corrected after more than one year. Furthermore, since growth rates are based on sales, a large variation in this number may implicate the purchase or selling of certain business units. Growth of Nuon in year -1 for example is probably the result of a merger between Nuon and another utility company. This study did not corrected for potential acquisitions or sales so data may be distorted.

Proposition 2:

(32)

Sample case firms liquidity position was in the years prior to outsourcing significantly worse than their industry counterparts’ liquidity position. Based on the information available, one can say that managers needed the money to improve their liquidity or finance short term operating needs. On the other hand, sample case firms did not seem to be subject to higher levels of leverage and to higher levels of debt so they did not needs the cash for refinancing purposes.

These results may implicate a distinction between a firms need for cash: cash in order to finance short term operating needs or cash to restructure and organizations capital structure or repay outstanding debts. Results indicate support for the assertion that organizations need the cash from the outsourcing deal to improve on their liquidity position. However, results reject the assertion that organizations need the cash for refinancing or debt repayment purposes. An explanation for these results may lie in the fact that higher debt ratios do not necessarily indicate a need to reduce debt and generate cash. Higher debt ratios can be profitable for firms when they can make use of potential tax shields (assumed no extra risks accompanied by higher debt levels exist). In this case, refinancing the organizations will only decrease its net present value. Hence, the assumption that higher debt levels indicate a firm’s need for cash may be untrue. What’s more, in the case of KPN for example which have low debt levels as one of their main objectives, relative debt levels do not give much information. When debt levels are already low (indicated by a negative sign in the results), there still may exist some desire to even further lower debt levels. Outsourcing part of the organization to obtain cash to finance this can still be a realistic option. So in this case it is still possible for KPN to outsource IS to obtain cash for refinancing purposes.

All in all, supports exists for companies needing cash to finance short term operations. However, more long term oriented goals such as restructuring the finance or debt repayments do not seem to hold.

Proposition 3:

(33)

indirect measures which comprehend an organizations’ measure of virtuality. Virtuality in turn is a term to describe the degree of focus on core activities an organization has. The degree of virtuality of an organization is measured in terms of number of employees they need to generate a certain amount of sales or profits. It is also measured as the amount of assets they need to generate a certain level of profit. The more assets or employees an organization needs, the less virtual it is because it is theorized that the organization buys less goods or services on the external market and produces more internally. Smith et al. (1998) illustrates this point with a comparison of Coca-Cola Company and PepsiCo which both operate in the same industry. In the 1988-1993 period, Coca-Cola had an average sales/employee value of approximately $600.000. PepsiCo however, had an average sales/employee value of only $65.000. The difference in value for both companies was hardly explainable by higher productivity of Coca-Cola company employees. The case here is that Coca-Cola company outsourced its bottling operations to a separate company and retaining only its core activity (marketing and manufacturing of the syrup) in-house. PepsiCo has its own bottling operations and consequently needs more employees / assets for similar levels of revenues (which is only generated by the selling of cola).

When examining the results of the empirical study conducted here, some interesting results become obvious. On each of the metric discussed, sample case firms score lower than their industry counterparts, and in most cases statistically significant lower as well. These results indicate that the sample case firms are less focused on their core activities than their industry equivalents. This might indicate good motives for managers to outsource their IS: they want to focus the organization just as other organizations are doing. So, IS provision, which was not a core activity for the sample case firms, is outsourced as a step towards the direction of a more focused organization.

(34)

Proposition 4:

The final proposition states that firms that outsource their IS do so because they are relative underperformers in their industry. It is asserted that firms who outsource their IS are less profitable and efficient. Managers can intend to increase short term profits or send a positive message to their shareholders by choosing to outsource their IS. This proposition is directly testable by publicly available financial data.

The results of this empirical study are not conclusive. Sample case firms do not seem to differ from their industry counterparts with respect to their efficiency and their return on equity. On the other metrics however, sample case firms did under perform their industry counterparts. Profit margins were lower every year for the sample case firms, and statistically significant in one year. Return on assets was also lower for the sample case firms in every year. This difference was statistically significant in two years. There also appear to be some significant trends in these metrics, however these are not unambiguous, so no real conclusions can be drawn from these trends.

Off course, it can be the case that at the time of outsourcing the whole industry was in a slump. In this case managers may still have had bad performance as a motivation for outsourcing their IS. However, since the whole industry is in a downturn, relative underperformance is masked and cannot be uncovered by the dataset. This is an inherent disadvantage of using these metrics and cannot be overcome by making otherwise use of publicly available financial data.

On the base of the available results, the fourth proposition cannot be accepted. Dutch based companies who outsourced their IS do not seem to persistently under perform the market. Though some evidence points in this direction, it is not compelling enough to have the proposition accepted.

(35)

4.2 Empirical Comparison

The study conducted here is based on a benchmark study conducted in the United States by Smith et al. (1998). This section will analyze the results found in this study and compare both results. In table 6 a summary is given of the major results of both studies.

The study conducted by Smith et al. incorporated IS outsourcing cases which took place in the period 1985 to 1995 in the United states. Sample case selection was comparable with the selection method in this study. A final sample of 28 firms was created. Propositions which were put forward were also similar to the propositions discussed here, with however a few alterations which will be discussed later. These propositions in turn were studied by measuring again similar financial characteristics of sample firms and comparing them with their industry counterparts.

Smith et al. found strong support for the proposition that firms outsource their IS to achieve cost cuttings. Though metrics were measured roughly likewise as in this study, a basic premise underlying the proposition was different. Smith assumed that organizations that outsourced IS were more cost conscious than other firms, and other cost cutting measures were already established.

Table 6: Empirical results comparison

Netherlands United States

Cost cuttings Not supported Supported

Cash generation Partly Supported Supported

Focus Supported Not supported

Low performance Not supported Not supported

(36)

IS, when only publicly available financial data is used. Furthermore, Smith et al. based their conclusions on another financial metric: dividends payout. This metric was significant for the results from Smith et al., however this metric was not incorporated in this study for data availability reasons. This may distort comparison of the results as well.

The second proposition is completely supported by Smith, however, the proposition is only partly supported in this study. Differences in the results lie in the fact that Dutch sample case firms did not differ from their industry counterparts with respect to debt levels and financial leverage while US firms did differ. This difference could be due to the fact that US firms are making more use of tax shields for example. It could also be that US firms do not have low debt levels as a priority (risk loving) while Dutch based firms do have low debt levels as a goal (risk averse). Again, since the metrics only provide indirect evidence, a difference in preferences between US organizations and Dutch organizations cannot be established without leaving some room for argument.

The third proposition, companies outsourcing IS to regain focus, is not supported by Smith but is supported in this study. This is an interesting dissimilarity to look into. Can it be the case that managers in the Netherlands tend to focus their companies more on core activities than managers in the United States? The difference can also be explained by the different periods in time in which the studies took place. Maybe businesses nowadays tend to focus more on their core activities and are more likely to outsource their non-core activities than the research window from Smith et al. (1985 – 1995). An interesting case which is beyond the scope of this research but worthwhile looking into for further research.

Determinants of the decision to outsource information systems in the Netherlands.