uncertainty (weak property rights) or bureaucracy,. 3 then unproductive ..... codes, as well as references to past literature, and/or solicitation of expert opinion to.
OUTSOURCING AS A FIRM GOVERNANCE MECHANISM: EMPIRICAL STUDY OF INFORMATION TECHNOLOGY OUTSOURCING DETERMINANTS AND PERFORMANCE, 1990-1999
APPROVED BY SUPERVISORY COMMITTEE:
_________________________________________ Euel Elliott, Co-Chair
_________________________________________ Donald A. Hicks, Co-Chair
_________________________________________ L. Douglas Kiel
_________________________________________ Sheila Amin Gutiérrez de Piñeres
Copyright 2000 Thomas N. Tunstall All Rights Reserved
To my wife Renée, and to Matthew, Rachel, Taylor and John
OUTSOURCING AS A FIRM GOVERNANCE MECHANISM: EMPIRICAL STUDY OF INFORMATION TECHNOLOGY OUTSOURCING DETERMINANTS AND RESULTANT OUTSOURCING PERFORMANCE, 1990-1999
by
THOMAS N. TUNSTALL, M.B.A.
DISSERTATION Presented to the Faculty of The University of Texas at Dallas in Partial Fulfillment of the Requirements for the Degree of
DOCTOR OF PHILOSOPHY IN POLITICAL ECONOMY
THE UNIVERSITY OF TEXAS AT DALLAS August 2000
ACKNOWLEDGEMENTS
I would like to express my profound appreciation to all of the members of my dissertation committee – Professors Euel Elliott, Donald A. Hicks, L. Douglas Kiel and Sheila Amin Gutiérrez de Piñeres. I am especially grateful to my Co-Chairs, both of whom I studied under during my first semester in the Ph.D. program in 1995. Dr. Euel Elliott volunteered to serve and help assemble a committee, and Dr. Don Hicks provided invaluable assistance with regard to both dataset acquisition and methodological rigor. I would like to thank Professor Wim P. M. Vijverberg and Paul Jargowsky for their assistance with data development for some of the independent variables. I am also grateful to Dr. L. Doulas Kiel and Dr. Sheila Amin Gutiérrez de Piñeres for their insight and guidance in the development of a suitable topic for research. This dissertation could not have been completed without the encouragement and support of my family, especially my wife Renée. As a part-time student with a full-time job and family, my challenge to balance work, home and school was made infinitely easier because of their efforts. My sincere thanks to Matthew, Rachel, Taylor and Jack for understanding that Dad sometimes had homework too. But in particular, I would like to thank Renée for providing the glue that kept the household together during what surely must have seemed an interminable period, that is at last coming to fruition.
v
ABSTRACT
OUTSOURCING AS A FIRM GOVERNANCE MECHANISM: EMPIRICAL STUDY OF INFORMATION TECHNOLOGY OUTSOURCING DETERMINANTS AND PERFORMANCE, 1990-1999
Publication No. ___________
Thomas Tunstall, Ph.D. The University of Texas at Dallas, 2000
Supervising Professors: Euel Elliott and Donald A. Hicks This study explores the determinants of firm IT (Information Technology) outsourcing and examines some of the early returns of the impact of IT outsourcing on firm performance. As with the rise of the multi-divisional form of corporate governance instituted by Alfred Sloan at General Motors in the 1920s and 30s, firms today are experimenting with outsourcing as a means to operate more effectively in a dynamic marketplace. In effect they are stripping away many in-house functions and exposing them to market forces. Since within-industry wage rates are comparatively homogeneous, wage rates across functions, but within industries, tend to be out of equilibrium with the market. This in turn can cause firms to have above-average costs and thus impede their competitiveness. Outsourcing is one tool firms have chosen to bring wage rates and costs into equilibrium. The study examines questions centering on why firms outsource using a data gathered from the 1990-1999 timeframe. The dataset consists of 299 outsourcing events vi
which have been combined with financial data reported prior to the event. Several conclusions can be drawn from the analysis. First, evidence strongly indicates that IT outsourcing by firms has increased substantially during the past decade. More importantly, firms engage in outsourcing as a way to decrease overhead costs (sales, general and administrative – often referred to in corporate parlance as cost centers). In addition, two industry groups, banking/financial services and transportation firms have higher outsourcing intensity than firms in other industry groups, at least in part as a function of their use of centralized systems, which are generally easier to outsource than distributed information systems. In addition, there is strong evidence to indicate that firms are entering into IT outsourcing contracts for shorter periods of time apparently to avoid becoming hostage to the outsourcer and maintain competitive pressures. There is also evidence of a shift in the types of IT functions outsourced over the decade, which may simply suggest that as the emphasis on and use of certain technologies shift, so does their tendency to be outsourced. A preliminary look at the effects of outsourcing on firm performance using the available subset of the 299 companies in the original dataset is inconclusive. There is no clear evidence that IT outsourcing alone impacts firm performance as measured by revenue per employee, asset efficiency, return on investment, return on sales, market-tobook ratio and revenue growth. As additional data is accumulated on outsourcing by research firms such as INPUT, more research should be performed in this area in order to gauge the extent that outsourcing has delivered on its promise.
vii
TABLE OF CONTENTS
ACKNOWLEDGEMENTS .........................................................................................................................V ABSTRACT ................................................................................................................................................ VI TABLE OF CONTENTS ........................................................................................................................ VIII LIST OF TABLES ........................................................................................................................................X LIST OF FIGURES .................................................................................................................................... XI CHAPTER ONE ........................................................................................................................................... 1 INTRODUCTION .................................................................................................................................... 1 1.1 MARKET VS. HIERARCHY (TRANSACTION VS. AGENCY COSTS)........................................................ 4 1.2 CHRONOLOGY OF OUTSOURCING ..................................................................................................... 10 1.3 OUTSOURCING AS A GOVERNANCE MECHANISM ............................................................................. 14 CHAPTER TWO ........................................................................................................................................ 18 THEORETICAL BACKGROUND AND EMPIRICAL EVIDENCE ................................................ 18 2.1 THEORETICAL PERSPECTIVES .......................................................................................................... 18 2.1.1 Previous Research on the Decision to Outsource ...................................................................... 19 2.1.2 Management Theories Regarding Outsourcing ........................................................................ 20 2.1.3 Practical Application on an Experiential Basis ......................................................................... 21 2.1.4 Case Studies ................................................................................................................................ 24 2.1.5 Representative Samples .............................................................................................................. 27 2.1.6 More Elaborate Models .............................................................................................................. 33 2.1.7 Opportunities for Extending Existing Work .............................................................................. 36 Dependent Variable: Individual Outsourcing Agreements ................................................................ 41 Dependent Variable: Degree of Outsourcing by Firms ...................................................................... 41 CHAPTER THREE .................................................................................................................................... 45 RESEARCH DESIGN AND MODEL DEVELOPMENT ................................................................... 45 3.1 RESEARCH QUESTIONS ...................................................................................................................... 45 3.2 DATABASE DEVELOPMENT AND CONCEPT MEASUREMENT ............................................................ 46 3.2.1 Unit of Analysis .......................................................................................................................... 46 3.2.2 Firm Categorization ................................................................................................................... 46 3.3 DETERMINANTS OF OUTSOURCING ................................................................................................... 48 3.3.1 Dependent Variable .................................................................................................................... 49 3.3.2 Predictors and Covariates........................................................................................................... 52 3.3.2.1 Profitability .............................................................................................................................. 52 3.3.2.2 Company Size........................................................................................................................... 52 3.3.2.3 Growth...................................................................................................................................... 53 3.3.2.4 Cost Focus................................................................................................................................ 53 3.3.2.5 Cash Flow ................................................................................................................................ 54 3.3.2.6 Organizational Focus .............................................................................................................. 54 3.3.2.7 Industry Effects........................................................................................................................ 56 3.3.3 Summary of Variables ................................................................................................................ 56 3.3.4 Outsourcing Determinants ......................................................................................................... 61 3.4 RESEARCH DESIGN AND DATA DEVELOPMENT ISSUES .................................................................... 62 3.4.1 Quality of the Data ..................................................................................................................... 63 3.5 INPUT DATABASE – SUMMARY STATISTICS .................................................................................... 64
viii
3.6 IT OUTSOURCING DETERMINANTS ................................................................................................... 65 3.6.1 Summary Statistics – Determinants Dataset .............................................................................. 66 CHAPTER FOUR ...................................................................................................................................... 70 OUTSOURCING AND FIRM GOVERNANCE .................................................................................. 70 4.1 RESULTS OF REGRESSIONS – DETERMINANTS OF IT OUTSOURCING .............................................. 70 4.2 INDUSTRY EFFECTS............................................................................................................................ 79 4.3 OUTSOURCING EFFECTS .................................................................................................................... 79 4.4 OTHER FINDINGS ............................................................................................................................... 80 CHAPTER FIVE ........................................................................................................................................ 85 SUMMARY AND POLICY IMPLICATIONS .................................................................................... 85 5.1 RESEARCH FINDINGS ......................................................................................................................... 85 5.2 PUBLIC POLICY IMPLICATIONS ......................................................................................................... 86 5.3 SUGGESTIONS FOR FUTURE RESEARCH ............................................................................................ 92 APPENDICES............................................................................................................................................. 94 APPENDIX A – IT OUTSOURCING DETERMINANTS SAMPLE STATISTICS .............................................. 95 APPENDIX B – OUTSOURCING EFFECTS METHODOLOGICAL BACKGROUND ....................................... 96 SUGGESTIONS FOR FUTURE RESEARCH – OUTSOURCING EFFECTS ...................................................... 96 Dependent Variable ............................................................................................................................. 96 Predictors and Covariates.................................................................................................................... 97 Degree of Outsourcing ........................................................................................................................ 98 Cost Control (Quality of Management) .............................................................................................. 98 Cash Flow (Quality of Management) ................................................................................................. 99 Experiential Effects and Dummy Variables ....................................................................................... 99 Summary of Variables ....................................................................................................................... 100 Outsourcing Effects Variables .......................................................................................................... 101 Causal Diagram ................................................................................................................................. 102 Outsourcing Effects ........................................................................................................................... 102 REFERENCES ......................................................................................................................................... 104 VITA .......................................................................................................................................................... 111
ix
LIST OF TABLES
TABLE 1.1.1 – BREAKDOWN OF SERVICES OUTSOURCED ............................................................................ 9 TABLE 2.1.1A - ANALYSIS OF DEPENDENT VARIABLE ............................................................................... 41 TABLE 2.1.1B - ANALYSIS OF VARIABLES AND DIAGNOSTIC RESULTS ..................................................... 41 TABLE 2.1.2 - ANALYSIS OF VARIABLES – INDEPENDENT VARIABLES ...................................................... 42 TABLE 3.3.1 – OUTSOURCING FUNCTION DEFINITIONS ............................................................................. 51 TABLE 3.3.1 – INDEPENDENT AND DEPENDENT VARIABLES ...................................................................... 58 TABLE 3.3.2 – CORRELATION MATRIX - MEANS ....................................................................................... 59 TABLE 3.3.3 – CORRELATION MATRIX – LAGGED 3 YEARS ...................................................................... 59 TABLE 3.3.4 – CORRELATION MATRIX – LAGGED 2 YEARS ...................................................................... 60 TABLE 3.3.5 – CORRELATION MATRIX – LAGGED 1 YEAR........................................................................ 60 TABLE 3.5.1 – FIRMS VS. NON-FIRMS ......................................................................................................... 64 TABLE 3.5.2 – NON-FIRM BREAKDOWN BY TYPE OF ENTITY ................................................................... 64 TABLE 3.5.3 – BREAKDOWN OF IT FUNCTIONS OUTSOURCED – ALL ENTITIES ....................................... 65 TABLE 3.5.4 – BREAKDOWN OF IT FUNCTIONS OUTSOURCED: FIRMS ..................................................... 65 TABLE 3.6.1 – BREAKDOWN BY SIC CODE ................................................................................................. 66 TABLE 3.6.2 – BREAKDOWN OF TECHNOLOGY FUNCTIONS OUTSOURCED .............................................. 67 TABLE 3.6.3 – DETERMINANTS SAMPLE OF ANNUAL IT OUTSOURCING BY YEAR ................................... 67 TABLE 3.6.4 – OUTSOURCING TREND AND FORECAST STATISTICS BY RESEARCH FIRMS ....................... 68 TABLE 3.6.5 – GROWTH OF OUTSOURCING FIRMS OVER THE PAST 20 YEARS ........................................ 69 TABLE 4.1.1 – DEPENDENT VARIABLE OIIT - DETERMINANTS OF IT OUTSOUCING ............................... 76 TABLE 4.1.2 – DEPENDENT VARIABLE OIAT - DETERMINANTS OF IT OUTSOUCING .............................. 77 TABLE 4.1.3 – DEPENDENT VARIABLE OIREV - DETERMINANTS OF IT OUTSOUCING ........................... 78 TABLE 4.4.1 – IT CONTRACT TERM ........................................................................................................... 82 TABLE 4.4.2 – IT CONTRACT DATE ............................................................................................................ 84
x
LIST OF FIGURES
FIGURE 1.1.1 – MARKET-HIERARCHY CONTINUUM .................................................................................... 8 FIGURE 2.1.1 – OUTSOURCING LIFE CYCLE............................................................................................... 19 FIGURE 3.1.1 -- CAUSAL DIAGRAM ............................................................................................................. 48 FIGURE 3.6.1 – DETERMINANTS SAMPLE OF ANNUAL IT OUTSOURCING BY YEAR ................................. 67
xi
CHAPTER ONE INTRODUCTION
Transactions – social, political, economic – are the means by which we seek to maximize value and utility. Whether attending a party to meet or interact with others, or using the voting mechanism to express ideological or collective preferences to maximum effect, we make use of transactions to enhance utility in any number of ways. Within the economic realm, there are two broad classifications of transactions. The most familiar of these, largely because they are the most easily measured, are those transactions exposed to the marketplace and subject to all corresponding constraints. Less obvious are those transactions insulated from market forces either because of social institutions (e.g., family homemaking), legal constraints (prostitution, drugs, significant regulation), or because they are exchanged within the boundaries of the firm. Such transactions occur in what has been term non-market institutions, and include families, governments, regulated markets and firms. This last group of these non-market transactions that occur within firms have been the hallmark of the rise of the multinational firm in the first half of the 20th century. Firms vertically integrated because market transaction costs were relatively high or because there was a lack of available suppliers in the marketplace for a given function. More recently however, particularly over the past 10-20 years, it is less clear that bringing transactions within the boundary of the firm creates economies. Given widespread deregulation starting around 1978, globalization of industries and falling trade barriers, and increasingly mobile labor and capital, the costs of economic transactions continues to come down. \
1
Many of the arguments that apply to globalization could be pertinent to outsourcing writ large:
Globalization clearly benefits producers by giving them greater choice over their raw materials, production techniques, and human talent, not to mention over the markets where they sell the goods. Equally clearly, globalization benefits consumers by providing them with better goods at better prices. Globalization increases efficiency and thus prosperity (Micklethwait and Wooldridge, 2000:335).
This dissertation will explore the proposition that because of decreasing transaction costs, firms are actively exposing operational functions to direct market forces driven by the need to increase economies, and in the process of doing so, making more efficient use of the factors of production. In the same way that money increases the number of possible transactions by orders of magnitude and allows an economy to build wealth at a much faster pace than would otherwise be possible (Gordon, 2000), lower transaction costs also increase the number of possible transactions and correspondingly allows an economy to build wealth at a much faster pace than would otherwise be possible. In effect, outsourcing better matches market signals with the factors of production and enables greater economic efficiency. A central question pertaining to the structure of the firms is why do managers choose to place the performance of certain functions outside the boundaries of the firm? According to Ronald Coase, they will do so only if it is more economical to do so, including explicitly the cost of transactions. This was the significant contribution made to neoclassical economics, which had previous assumed frictionless markets. But markets are not frictionless, and the method or mechanism by which firms choose to
3
organize their production activities, i.e. their governance mechanism, matters very much on a variety of levels. The key distinctions that will be made in this paper are twofold: the mechanisms by which firms govern either (1) within their boundaries, or (2) alternatively, across them. Governance across the boundaries of the firm is a mechanism also commonly referred to as outsourcing. Not only is the overall cost structure of the firm potentially impacted by the governance mechanism – in this case the decision to outsource – so are the characteristics of control or power utilized by management, and the degree of specialization exercised by different functional areas. While it is sometimes argued that outsourcing will tend to reduce the average size of firms (Downes and Mui, 1998), such statements are misleading. Outsourcing does not reduce the average size of the firm, at least when measured in revenue 1, since one company’s expense is another company’s revenue. An outsourced activity shows up as cost of good sold or general overhead on the firms profit and loss statement, the same as it would if it were performed and managed in-house. Revenues must be in excess of costs, which include outsourcing, in order for a firm to be profitable. Outsourcing simply increases the number of market transactions, exchanges, or interactions that occur during the process of taking a product or service to market. The result of decreasing transaction costs and the rise of the virtual corporation then is that outsourcing changes the nature of firm boundaries and establishes more of them. Interestingly, while power exercised by firms within their boundaries is presumed to confer greater control, the evidence in this study suggests otherwise. In fact the evidence supports the thesis that market competition enables firms to achieve greater 1
This is the criterion that the Fortune and Forbes 500 use for their respective rankings of firm size.
4
control over their costs, and that outsourcing exposes above-market costs for factors of production, particularly labor (as a principle component of services) in a way that internal bureaucratic control cannot. The increased use of outsourcing during the 1990s shows that firm boundaries are becoming increasingly permeable. Further, greater numbers of transactions would seem to imply lower transaction costs in an economic system. Much of these changes, like those associated with the Internet, suggest better supply chain linkages (i.e., the ability to choose from a multitude of suppliers) and the release of factors of production to their most productive use by unbundling, decoupling and redeploying them.
1.1 Market vs. Hierarchy (Transaction vs. Agency Costs) Ronald Coase (1937) first postulated that firms were organized primarily to reduce transaction costs such as search, information, contracting, monitoring, and enforcement costs 2. According to Coase, firms will expand their boundaries as long as agency costs are less than transaction costs. Agency costs are typically defined as transaction costs which occur inside the firm. Whiles transaction costs are incurred when two market players interact and perform an exchange, agency costs are within-firm or non-market exchanges. Technology is enabling more transparent and rapid communication between firms. With increased complexity and specialization, new forms of organization will be the subject of experimentation in an effort to lower costs or drive growth, as Coase indicated:
The question always is, will it pay to bring an extra exchange 2
Transaction costs are generally defined as search, information, bargaining, decision, policing and enforcement costs.
5
transaction under the organising authority? At the margin, the costs of organising within the firm will be equal either to the costs of organising in another firm or to the costs involved in leaving the transaction to be “organised” by the price mechanism. Business men will be constantly experimenting, controlling more or less, and in this way, equilibrium will be maintained. This gives the position of equilibrium for static analysis. But it is clear that the dynamic factors are also of considerable importance, and an investigation of the effect changes have on the cost of organising within the firm and on marketing costs generally will enable one to explain why firms get larger and smaller. We thus have a theory of moving equilibrium. The above analysis would also appear to have clarified the relationship between initiative or enterprise and management (Coase, 1937: 404-05).
History has shown us that firms are inclined to grow and expand. Technology and process are enablers of that inclination. Coase acknowledges this by stating that “all changes which improve managerial technique will tend to increase the size of the firm.” However, Coase also supplies a caveat: It should be noted that most inventions will change both the costs of organising and the costs of using the price mechanism. In such cases, whether the invention tends to make firms larger or smaller will depend on the relative effect on these two sets of costs. For instance, if the telephone reduces the costs of using the price mechanism more than it reduces the costs of organising, then it will have the effect of reducing the size of the firm (Coase, 1937).
Coase suggested that there was a cost associated with using the price mechanism, such as the search costs of discovering what the relevant prices are. And while many products lend themselves to spot market transactions where competitive prices are easy to discern, some products and even more services are more complicated. In order to economize on numerous individual exchanges, parties may opt to enter into longer-term contracts and thus avoid some transaction costs in the process. This is where the case for
6
the firm is made. Outsourcing enters the picture as sort of a middle ground between transactions amenable to the spot market and vertical integration (i.e., taking the entire activity or function in-house and thus expand the boundaries of the firm). It is here that the market (outsourcing) serves to increasingly erode traditional firm boundaries. Williamson has noted that successive levels of bureaucracy make organizations less efficient because of diffusion of communication. Coupled with differing goals at each level and the firm can become quite inefficient:
For any given span of control (together with a specification of the state of technology, internal experience, etc.) an irreducible minimum degree of control loss results from the simple serial reproduction distortion that occurs in communicating across successive hierarchical levels. If, in addition, goals differ between hierarchical levels, the loss in control can be more extensive (Williamson, 1986: 47).
While there appears to be little debate among management theorists as to why firms choose to use outsourcing as a governance mechanism, and thus choose the market for performance of a function or service instead the firm (hierarchy), much of the theorizing is not subject to systematic scrutiny. While economic presumptions are that the market gives preference to the form of organization that could be termed the "leastcost performance mechanism" over the long-term, survey evidence suggests that there are drivers other than cost reduction. These other drivers, such the need improve organizational focus, may suggest that outsourcing is evolving into an effective organizational management tool. Yet a clean definition of what has been termed core competency (Prahalad and Hamel, 1990) remains elusive, even potentially unique for every firm. As such, this defies any consistently measurable definition. Nonetheless, the
7
general theme of increased firm performance by management theorists is increased focus and greater specialization. Increased specialization is one way in which an economy increases overall output relative to input (Cheung, 1998; Deavers, 1997). It may be useful to think of outsourcing as an organizational tool by which a firm may increase its ability to specialize. Yet increased specialization, which is often enabled by outsourcing, leads necessarily to asset specificity of both labor and capital. Asset specificity can in turn lead to rent approbation by the outsourcer, leaving the firm with little leverage and subject to a hostage situation. Rent approbation is one of several issues that can arise where asset specificity is present and which can have both a macroeconomic and microeconomic impact (Caballero, 1998). In effect, the inability of economic concerns to efficiently allocate assets because of asset specificity, which are typically viewed from a microeconomic perspective, have macroeconomic implications as well. While some of these issues can be mitigated through contracting techniques; others are the result of the shifting balance of the power of the market as it pushes toward macroeconomic efficiency. Buffers that slow or impede this competition-driven push are found in the power relations among competing concerns and institutional framework, either at the state or industry level (Baker, et al., 1998). These include non-market institutions such as labor unions, federal agencies, state agencies, and as mentioned previously, excessive regulation. Outsourcing can strip away selected firm functions and remove the buffers/impediments to high-powered market signals. This in turn channels resources to their most productive use – or at least to relatively more productive uses. Closed organizational structures, whether countries or firms, become rigid and hierarchical, and ultimately resistant to innovation. Outsourcing
8
is usually a response to outmoded organizational structures that cannot transform quickly enough on their own. And outsourcing is in turn enabled by lower transaction costs. Other issues surrounding outsourcing involve trade policy and globalization, particularly with regard to wage levels as a function of the supply and demand of labor for different job functions, and economic restructuring, which is being enabled by new technologies. As a point of clarification, outsourcing has been variously defined to include both manufactured goods and services. Outsourcing has also been defined in terms of the nature of the governance relationship over the function being outsourced. At one extreme is a simple spot market (unilateral) contract for a tangible good; at the other end of the spectrum is the in-house performance (internal organization) of a function by the firm. Outsourcing will tend toward the market side of the continuum. Elements along the governance continuum have been outlined in Figure 1.1.1.
Figure 1.1.1 – Market-Hierarchy Continuum
Markets
Unilateral Contract
Bilateral Contract
Equity/ Joint Venture
Internal Organization
Hierarchies
Adapted from Oxley, 1997
While a very broad definition of outsourcing can include both goods and services, the data analyzed in this dissertation are focused on services. Further the data address only a subset of the total range of services that it is possible for firms to outsource. This limitation is due mainly to a paucity of available data. Data for goods producing sectors, by comparison, are generally more prevalent, and easier to obtain and track, in part because it is easier to measure tangible goods than it is services. Materials procurement,
9
for example, has been studied extensively and is in many ways a form of outsourcing. The domination of the service sector in advanced economies is relatively recent, and the data are, frankly, more difficult to measure and less well developed as a result. The categorization of services outsourced by firms is outlined in Table 1.1.1. As is evident, nearly all manner of services required by a firm can be outsourced if its chooses to do so.
Table 1.1.1 – Categorization of Services Outsourced Organizational Function Information Technology (IT) Distribution Logistics Real Estate Facilities Management Administration Human Resources Customer Service Finance Marketing and Sales Transportation Functions Total Services Outsourcing
Percent 40% 30%
30%
100%
Source: Outsourcing Institute, Jerico, NY (www.outsourcing.com)
Perhaps the most detailed categorization of IT services is available from Clark, Zmud and Gray (1998):
Data center installation/operations/maintenance Data storage Facilities management Hardward and software maintenance LAN support Network design/operations/maintenance/control (voice/data) PC support Professional services (consulting, planning, technology assessment) Programming services Remote operations Security and disaster recovery
10
Service center (transaction processing or MIPS utility) Software development process (any or all phases) Staff and/or user training Systems installation and integration Systems software
As can be noted, there are clearly possible overlaps of definitions between IT and non-IT services. In addition, other authors provide different epistemologies regarding the categorization of IT functions which will be discussed in the next chapter. This dissertation will examine determinants of IT outsourcing as a governance mechanism in order to identify or validate causal links as to the degree of outsourcing a firm may undertake. In addition, the data are amenable to exploring a limited number of results from outsourcing, again in a comparatively objective and systematic fashion, which I will detail in the analysis that follows.
1.2 Chronology of Outsourcing By way of background, IT outsourcing traces its origins at least as far back as the 1960’s, and perhaps even before that (e.g. ADP payroll processing), as firms began to look at outsourcing as a means to more effectively manage particular functions. Arguably, the first instance of IT outsourcing was for the tabulation of the data collected during the 1890 census by Herman Hollerith and his card punch/reader system (Lee, 1995: 378-79). Generally, however, IT outsourcing is acknowledged to have begun when Ross Perot unsuccessfully attempted to convince IBM to create a computer services division to manage data processing facilities for corporations. Perot then left IBM to form EDS in 1962, and offered to run data centers for companies faster and cheaper than they could on their own (Levin, 1989: 24-29). Such a scenario is often indicative of
11
otherwise bright and innovative companies who fail to see the future. IBM has since aggressively entered the services market and is one of EDS’ prime outsourcing competitors. Interestingly, the company that Herman Hollerith formed (Tabulating Machine Company) was eventually merged with three other companies in 1911, at first called the C-T-R Company, and became known as IBM in 1924 (Austrian, 1982: 311-12). In part, the move to outsourcing was driven by managerial ineffectiveness in noncore areas and/or as a way for distressed firms to free up cash flow by making capital investment in computers (IT) more liquid. Recently however, particularly since Kodak outsourced large portions of its IT function to IBM, effectively legitimizing the strategy, outsourcing has become an accepted means for management to more efficiently run operations (Loh and Venkatraman, 1992). Organizations have been through such transformations in the past. The relatively recent development of large-scale industrial hierarchies in the past 70-80 years is probably the most notable example. When comparatively rigid, top-down organizational structures became inflexible and inefficient, the opportunity to innovate was exploited. The resultant M-form or multiple division organization, which was first pioneered by Alfred Sloan at General Motors and Pierre S. du Pont at Du Pont in the 1920’s, is an example of such organizational innovation. The M-form has proven an effective means for firms to manage agency costs by creating semi-autonomous operational units, yet retaining a central staff to monitor performance, allocate resources among divisions and engage in strategic planning (Williamson 1986). Prior to the M-form innovation, large firms with more centralized, top-down hierarchies (U-form or unitary form organization) had previously been the means by which organizations managed their operations.
12
Limitations of the U-form organization include bounded rationality constraints of upper management as well as the diffusion of market signals through successive layers or hierarchy. "Bounded rationality", a concept introduced by Simon (1976), acknowledges the cognitive limitations of managers. As an organization grows, the ability of any one person to manage in a detailed fashion diminishes, while the need for some type of control often remains. Similarly, as the organization grows, the market signals that discipline firms through competition become diffused, and manifest themselves in ways which decrease the efficiency within a hierarchy. These factors include excessive bureaucracy, internal politics, a lack of common objectives, and the use of firm resources for personal reasons, which Leibenstein (1966) identifies as X-efficiency factors. These X-efficiency factors contrast with the more traditional economic focus on allocative efficiency or the most efficient use of the factors of production (i.e., social welfare costs of monopoly; international trade restrictions). Bounded rationality and diffused market signals by agents on the one hand, with the simultaneous need to increase profits and drive shareholder value by principals on the other, are the conflicting forces that drive organizational change. Williamson, for example, highlights the plight of the senior executive in a U-form organization: …he can acquire additional information only by sacrificing some of the detail provided to him previously. Put differently, he trades off breadth for depth in undertaking the expansion; he has more resources under his control, but the quality (serial reproduction loss) and the quantity (bounded capacity constraint) of his information are both less with respect to the
13
deployment of each resource unit. In a similar way, being further removed from the operational situation and have more subordinates means that his instructions to each are less detailed and are passed across an additional hierarchical level (Williamson 1986, p. 36).
Outsourcing may represent a form of organizational innovation that will have as much impact as the M-form has had over the past seventy years. Since outsourcing is a recent organizational innovation, the literature is relatively young and still developing. IT outsourcing in its current form began in 1963 when EDS began providing limited data process services for Frito Lay and Blue Cross & Blue Shield (Teng, Cheon, Grover, 1995). However, the widespread acceptance of IT outsourcing has only been legitimized in the past decade or so. The mimicking behavior of corporate management after the Kodak deal with IBM appears to have had significant diffusion effects in terms of legitimizing outsourcing. Prior to the Kodak effect (Loh and Venkatramen, 1992), outsourcing was viewed more as a tactic for firms in trouble or with cash flow problems. These dynamics appear to be at least part of the reason that companies are increasing their use of outsourcing as an organizational management tool. Possible reasons for the increase in outsourcing stem from organizational inefficiencies that occur either through excessive agency costs or the advantages accruing to specialization and scale economies. There is growing recent evidence however that firms increasingly view outsourcing as a proactive strategy to enable growth, much like the innovations introduced at GM by Alfred Sloan in the 1920’s and 30s. In addition to the work done by Chandler (1962), Norton (1997) also examined the corporate governance mechanism at GM. He found that improved forecasting and information flow with the hierarchy (multi-
14
divisional corporation) was also a significant determinant of the success of General Motors in the 1920’s and 1930’s.
1.3 Outsourcing as a Governance Mechanism Firm inefficiency can occur for a variety of reasons. One of the causes stems from the ability of managers and employees to mask the true costs of coordinating and performing functions, an outgrowth of information asymmetry (Leibenstein, 1966). As in-depth knowledge by workers increases their task expertise, the benefits may not necessarily be passed on to the firm. Instead, workers may choose to make use of their knowledge to accomplish personal objectives instead of increasing organizational efficiency, particularly if those objectives are not sufficiently aligned with the firm’s goals. To a large extent, the ability of a firm’s employees to obscure agency costs is driven by the complexity of the function involved. IT functions are inherently more complex, for example, than facilities maintenance (e.g., landscape services), which are more widely understood in terms of definition and scope, use lower-skilled labor and are subject to greater market competition. As a result, there are probably more significant opportunities for inefficiency to be masked within the IT function. By the same token, there are reasons, however, why keeping functions within the firm may make sense outside strict scale economy advantages. Complex ownership relations, for example, makes contracting more difficult. Explicit contracts increase relationship rigidity more than implicit contracts. So clearly outsourcing, which relies on contracts instead of hierarchical control, is not without risks, particularly with regard to long-term contracting and the associated hazards of appropriable quasi-rents (Klein,
15
Crawford, Alchian, 1978). In effect, while agency costs can be high, so can the costs associated with asset specificity and opportunism by outsourcers. Nonetheless, given that principals (e.g. shareholders, senior managers) do not immediately act when agency costs are higher than transaction costs, two scenarios tend to play out as a result. Either a firm will at some point adopt a more innovative organizational form, or it will eventually be displaced by others which operate more efficiently, as suggested by the Schumpeterian notion of creative destruction. If government regulation makes contracting a more expensive alternative because of uncertainty (weak property rights) or bureaucracy, 3 then unproductive firm hierarchy will result. As pressure builds, buffered in the short-term by protection from direct market signals, the market eventually exerts its influence through less direct means such as capital markets, management labor markets (Fama 1980), new firm entrants and outsourcing proposals to senior management. These indirect influences provide discipline in the following ways: 1. Capital Markets: Investors and investor groups (mutual funds, venture capitalists) insist on competitive rates of return and, particularly in the past decade, have put increasing pressure on firms to outperform comparatively less risky investment vehicles such as government and corporate bonds. 2. Labor Markets: Managers who can demonstrate positive earnings results can reap significant rewards, which in turn drive more market-oriented behavior on their part, encouraging them to act more as principals than as agents. 3
The equivalent of internal agency costs.
16
3. New Entrants: These can take the form of direct competitors who enter a market enticed by the above-market returns of incumbent firms, or as substitutes, often abetted by lower cost models. These factors keep pressure on the incumbent firms to maintain competitive cost structures. 4. Outsourcing Proposals: These may arrive solicited or unsolicited, and can be used as a tool for senior management to more rapidly implement organizational change, improve performance, generate cash and/or reduce costs.
Outsourcing emphasizes existing cleavages along functional lines and is another Schumpeterian disruptor that challenges old hierarchies and ossified business models. Given the relative intractability of minimizing agency costs in a firm as it grows larger, a comparatively low transaction cost environment should serve to increase economic wellbeing by insuring competition with intra-firm transaction costs (agency costs). It is in this way that agency costs have the greatest likelihood of being minimized, and that competition, flexibility and adaptive organizational response are maximized. In such an environment, firm size will then be governed by compelling coordination needs, complexity of tasks, or scale economies – in short, activities that improve efficiency and economic surplus. Hence, in a relatively open and unregulated market, benefits accruing to producer specialization and comparative advantages (minimization of opportunity costs) lead to increased overall economic well being because they increase efficiency and choice. This perspective, outlined by Adam Smith (1776) and David Ricardo (1817), is the basis for the bulk of our material prosperity today. Otherwise we might well be consigned to
17
growing our own food, making our own clothes, or more improbably, building our own cars and manufacturing our own integrated circuits. Lower transaction costs encourages both greater specialization and increased interaction/trade among the factors of production. Lower transactions costs also enables the use of outsourcing, which may in turn increase consumer and producer surplus.
CHAPTER TWO THEORETICAL BACKGROUND AND EMPIRICAL EVIDENCE 2.1 Theoretical Perspectives Engaging in an IT outsourcing agreement by a firm is more complex than the simple procurement of goods. Instead of the manufacture of a product or good that is easily measured, defined and inspected, IT outsourcing generally refers to the transition of processes and services to a third-party, which are comparatively more difficult to measure and quantify. In order for a firm to outsource, it must first make the decision to do so, or at least undertake a serious examination of the prospect. It must then contact vendors, often through a request for proposal (RFP) process which outlines what services the firm wishes to obtain. Sometimes bidding ensues, although handshake deals without formal bidding have been relatively frequent, particularly in the early 1990s. Once a vendor is selected, the contract is negotiated and signed, and then the affected facilities, equipment and personnel are transitioned to the outsourcer. When the contract comes to term, the firm must then decide whether to renew with their existing outsourcer, re-bid the contract, or less frequently, bring the function back in-house. While the steps vary for each incident of outsourcing, the general process is outlined in Figure 2.2.1. Once negotiation is complete and the contract has been signed, the “fundamental transformation” has occurred (Williamson, 1985) in which the relationship is transformed from an ex ante competitive setting to an ex post bilateral monopoly. It is at this point that leverage can shift from the buyer to the seller of IT outsourcing services, depending on the nature of the contract.
18
19
Figure 2.1.1 – Outsourcing Life Cycle Decision to Outsource
Request for Proposal
Bid Process
Vendor Selection
Contract Negotiation
Transition
Exit/ Renewal
Source: KPMG LLC, Everest Group
From here we can proceed to examine the research that has gone before and what has been learned to date.
2.1.1 Previous Research on the Decision to Outsource Many industry associations, consulting firms and research firms have performed surveys as to why firms outsource. While the percentages vary, particularly since many of the results are from convenience samples, most of the same drivers continually emerge. Previous research in this area has focused mainly on case studies, interviews, and questionnaires to develop a comprehensive list of the key drivers of IT outsourcing (Smith, Mitra and Narasimhan 1998). Much less work has been on outsourcing using data from a large-sample publicly-held companies (Loh and Venkatraman 1992), which suggests opportunities for further research. This dissertation will examine IT outsourcing as a form of firm governance and test hypothesized determinants of outsourcing. It will explore the prospect that firms may be evolving into increasingly specialized, more focused organizations that grow either organically or through related horizontal acquisitions, as opposed to conglomerate-style horizontal acquisitions or vertical integration. Corporate evolution may be moving toward the use of full or partial spin-offs (divestiture) as a way to retain focus and reward principals. When new lines of business
20
become increasingly unrelated, as indicated by a different set of goals than that of the parent, firms may seek to create separate hierarchies to manage them. Internal transfers between the parent and the spin-off then become purchases and sales structured as market transactions (outsourced functions) between each other, as well as between other companies (England 1999). These developments imply lower transactions costs and more specialized firms, with outsourcing as the enabler.
2.1.2 Management Theories Regarding Outsourcing IT outsourcing has been approached from a variety of perspectives and with differing degrees of rigor. While many articles have relied on anecdotal evidence to suggest outsourcing’s role as an emerging corporate governance mechanism, they have nonetheless been useful in framing the discussion. For example, Quinn and Hilmer (1994), in what has become something of a seminal article for general management with regard to outsourcing, discuss how firm governance can be improved through identifying and focusing on core competencies based on attributes of human capital, as opposed to specific products and services that a company may provide to the marketplace. Outside of this core, it is recommended that hierarchies should be shunned in favor of the market for reasons of management focus (bounded rationality) and opportunity costs. Their discussion addresses organizational focus, which I attempt to operationalize in a number of ways that are described later. McFarlan and Nolan (1995) discuss how different levels of IT – operational to strategic – may affect the attractiveness of outsourcing and the nature of the relationship between seller and buyer. The divergent profit goals of customer and outsourcer are
21
examined. At the more strategic levels, companies may have increasing difficulty in finding a strategic fit with an outsourcer, and partnership (risk/reward sharing, as opposed to fee-for-service) may be a more effective governance mechanism. Venkatraman (1997) furthers the discussion by examining how IT can fall along a spectrum, ranging from a cost or operational activity, to a value-added activity. For value-added activities in particular, the IT function tends to permeate the business enterprise. Venkatraman and Henderson (1998) extend this argument by suggesting the need of greater sophistication of sourcing strategies, again from a theoretical perspective and aimed at practitioners. Still further, DiRomualdo and Gurbaxani (1998) recently examined outsourcing from a strategic standpoint, also targeted at practitioners. Each of these articles exemplifies potential methodological issues about how IT is defined, how it can permeate the firm, and where the boundaries are. These issues can be addressed in research, and several studies have developed relatively standard definitions of IT functional areas, which are generally accepted and are outlined later. The intent of the articles is primarily to articulate issues and perhaps suggest areas for research, but again, do not offer anything beyond anecdotal evidence to support their theories. They are, in effect, the jumping off point for more detailed analyses that will be reviewed in subsequent sections.
2.1.3 Practical Application on an Experiential Basis Along the same lines as management theory, some authors have discussed outsourcing with more of an eye toward providing the practitioner with useful advice, but also providing a sort of taxonomy for IT. These approaches are briefly outlined:
22
West (1994) in an article, which is not focused on outsourcing as a governance mechanism, nonetheless addresses the economics of information systems by functional area. West breaks down IT functional areas as follows:
Computation and processing power Software, operating systems and applications Data System design Data storage Telecommunications Overhead
He finds that in many cases, specialization by outsourcers can produce efficiencies in excess of transaction costs. While all of the functional areas had potential economies, there are differences in the magnitude of scale. West draws on the literature to support his conclusions, but does not offer any empirical evidence. Nonetheless, his is one of several taxonomies proffered for consideration to show, in part, that the IT function can be and is dissected in a variety of different ways. In a case study conducted by Lacity, Wilcocks and Feeny (1996) using 62 sourcing decisions in 40 organizations, outsourcing management practices were found to produce economies as well. Such research suggests that companies may outsource to lower costs, and this hypothesis will be included in the research model described below. The sample used by the authors was “purposive” in that they chose companies based on a wide range of situations -- from internal staffing providing the bulk of IT services to tenyear contracts where most or all functions were outsourced. The attempt was to be representative, but ultimately the sample is largely one of convenience. Still their study is one of the first attempts at an in-depth, systematic analysis of the use of outsourcing as a governance mechanism.
23
Earl (1996) discusses the risks of outsourcing IT and approaches the question from the perspective that insourcing (in-house performance of the IT function) is the null hypothesis. The article is intended for practitioners and identifies 11 risks of outsourcing. 1. 2. 3. 4. 5. 6.
In-house management problems may be exacerbated by outsourcing Personnel issues can arise during the transition phase Inflexibility of long-term contracts can be a disadvantage Outsourcers may employ older technologies Effective contracting may be more problematic than managing IT in-house Outsourcing customers may underestimate transition and ongoing management costs 7. Loss of organizational learning may occur 8. Loss of innovative capacity may stifle future firm growth 9. Outsourcing governance between users and suppliers can be problematic 10. Boundaries of responsibility between the outsourcer and the firm may be unclear 11. Firms who outsource may focus on the IT process instead of results and outcomes The author bases his recommendations presumably on his experience as a professor of information management and offers no empirical data or specific examples as support. No doubt many issues can arise for firms as a result of outsourcing. For one thing, the widespread use of outsourcing as form of governance, as a governance mechanism, is relatively recent. Inexperience with contracting (a form of transaction costs) in substitution for agency costs can indeed make governance problematic, and many firms have had difficulty with this transition. As a result, management theorists are quite right to point out some of these frequently encountered transitional issues that have occurred as outsourcing has increased. Further, it may be that because of such lessons, management has gotten collectively smarter about how it undertakes the use of outsourcing as a governance mechanism, and that more recently consummated outsourcing deals may be also more successful as a result. This hypothesis will be
24
explored later.
2.1.4 Case Studies Moving to the next level, but still very much with the practitioner in mind, the case study approach has been useful for clarifying issues and identifying potential variables for future research. Much case study work has been done by Lacity, Wilcocks and Feeny, who introduced the use of selective outsourcing and the fostering of competition among suppliers by not awarding everything to one provider in a single deal, as was done with Kodak. Their 1995 paper is one of several studies the authors undertook using the same dataset of 40 companies from a purposive sample in a variety of industries, including some in the public sector. The authors conducted 150 interviews for their practitioner-oriented case study. They found that many important IT functions were not necessarily strategic, and that managers often had difficulty identifying genuinely strategic functions. Also, many times the staff that previously managed the IT function before outsourcing were not able to make the leap to managing an outsourcing contract. In addition, ostensibly specific service levels were still subject to interpretation by the firm and outsourcer, which were driven at least in part because of divergent goals between the two parties. The outsourcer was focused on maximizing revenues and profits while the firm wanted lower costs and good service. Again, their dataset, while attempting to be representative must be considered a non-probability sample. Another case study that focused specifically on outsourcing was performed by Palvia (1995) who analyzed a single company using a dialectic approach. A dialectic is a debate between opposing viewpoints, and it is in the give-and-take of differing
25
perspectives that the “truth” is posited to emerge. In his study, Palvia attempted to obtain a balanced perspective from all of the stakeholders, not just comments from what he termed “the cheerleaders” (those customers and outsourcing company officials that tended to be quoted in the press). The case study dealt with a bank outsourcing deal where the participants in the process were interviewed. The study found that many of the outsourcer’s claims did not materialize, and the issue of divergent goals between the customer and outsourcer, which has appeared in previous research, was again raised. By the same token, there were significant problems in the form of agency costs associated with the internal IT organization, which were high (i.e., significantly above market). Also, because of the way the IT department was organized, the presumed hierarchical control within the firm was an illusion. While the study provides useful insight, a sample size of n = 1 may or may not be generalizeable to firm governance and outsourcing situations overall. If nothing else however, the Palvia study demonstrates potential issues and missteps that can occur as the by-product of an outsourcing deal, again suggesting that there is learning curve that may be associated with outsourcing as an organizational innovation. Segars and Grover (1995) perform a case study analysis of three industries: airlines, drug wholesaling and industrial chemicals. In each industry, IT innovations radically altered business conduct. Their most effective use, i.e., that which garnered the greatest competitive advantage, occurred in a centralized IT structure coupled with a geographically-dispersed operating structure. The successful firms used emerging network technologies to coordinate geographically dispersed operations and aligned the business structure with the IT systems. This created a competitive advantage not easy for
26
other firms in the industry to replicate. This effect was manifested in industrial chemicals and airlines, where the industry operating structures were comparatively heterogeneous. The operating structures of companies in the drug wholesaling industry by contrast were relatively homogeneous, and the effect of technological innovation was easily duplicated and thus short-lived. The study identified firms within the chosen industries using SIC codes, as well as references to past literature, and/or solicitation of expert opinion to resolve ambiguities. It did not specifically examine outsourcing, and its key limitation is that only three industries were studied. However the study does suggest potential differences between industries in regard to the IT function and the use of outsourcing, which will be examined further in the data analysis section. Cross, Earl and Sampler (1997) performed a practitioner-oriented single case study on outsourcing that found many positive elements associated with its use as a governance mechanism at British Petroleum. According to the study, British Petroleum was on the vanguard of firms seeking to use outsourcing as a means to both lower costs as well as implement organizational changes. The study cites three generations of literature on the IT function dating back to the 1960s. The first generation of IT was dedicated to lowering costs. The second generation of IT practice began to take hold in the late 1970s and 1980s, and was concerned with managing IT as a strategic resource. But because many of the presumed strategic benefits of IT did not materialize, the management of IT increasingly involved widespread experimentation, often times taking the form of outsourcing. The study does a good job of bringing to life many otherwise abstract notions regarding one firm’s experience with outsourcing. The dataset for this dissertation will enable analysis of such an evolution in the types of outsourcing occuring
27
only in a tangential way. Specifically this will entail examining the types of functions outsourced during the ten-year period to see if the distribution has changed over time. In still another study by Lacity et al., Lacity and Wilcocks (1998), using the same dataset in the two previous studies cited (Lacity, et al., 1995; 1996), re-analyzed transcribed interviews from the 145 participants. The authors came to several additional conclusions based on their analysis. Using perceived satisfaction of the respondents as the criterion, they found that selective outsourcing is more successful than total outsourcing. Outsourcing was most successful when there was coordination between senior executives and IT managers. Companies that secured both internal and external bids had the greatest success. Short-term contracts were more successful than long-term contracts. Finally, detailed fee-for-service agreements were more successful than comparatively vaguely worded fee-for-service agreements. One of these variables – the contract term – can be explicitly explored in this study in order to examine whether or not firms in fact are now opting for shorter-term outsourcing deals that was the case earlier in the decade.
2.1.5 Representative Samples As outsourcing issues have become increasingly well defined, many studies have begun to employ the use of surveys in order to obtain more representative results from larger populations. These studies draw their strength not upon opportunistic samples, but rather from sampling designs which that characterize a universe of firms. All of these types of studies are relatively recent as well. Saarinen and Vepsallainen (1994) surveyed 225 Finnish firms and obtained 55
28
usable responses. Results of the study indicated that the use of company-specific applications and high environmental uncertainty led to greater internal development of IT. The more standard the situation, the greater the use of outside consultants or software contractors. A possible source of bias cited is that Finnish firms use internal development more than has typically been the case in other countries such as the U.S. and U.K., and is likely the result of cultural predispositions. The chief limitations of this study are its use of data from a single, relatively small country, the fact that the scope was confined the applications areas, and the use of a comparatively small number of predictor variables. Arnett and Jones (1994) performed a survey of senior IS managers using a systematic random sample of 252 CIOs from the S&P Standard Register and obtained a 43% response rate. The study undertook measures of CEO involvement (CEO use of computers, CEO involvement on the IS steering committee), and the impact of that on the likelihood of outsourcing. The study also examined company profiles and found that those in leadership positions outsource the least, close followers outsource the most and laggards fall somewhere in between. CEOs who have heavy involvement in an information services (IS) steering committee are the least likely to outsource. CEOs that actively use computers are more likely to outsource specific hardware and software activities whereas CEOs who do not personally use a computer are more likely to outsource comprehensive management activities. In addition, the more levels there are between the IS manager and the CEO, the greater the likelihood that the IS functions will be outsourced. The authors surveyed for many independent variables which have not been widely used in the literature and cannot be obtained from public sources. Hence the results may be of limited use to practitioners and would likely be difficult to replicate in
29
future studies using alternative research designs. Teng, Cheon and Grover (1995) examine how as expectations rise and technology becomes increasingly complex, and as performance of in-house IT slips, outsourcing may become a response of strategic necessity. The two hypotheses which are supported by the study suggest that low quality of information supplied by IS and low quality of the IS staff/service in turn leads to a tendency to outsource. The two hypotheses dealing with the cost effectiveness of the IT function and overall financial performance of the enterprise as potential outsourcing drivers were not supported. The study used retrospective survey data for changes in the IT budget by function (present vs. recall of 3years previous). While the sample size is adequate and apparently representative, much of the data obtained depend on the ability of respondents to recall information and perceptions from three years previous, which raises reliability issues. This dissertation, by contrast, will rely exclusively on information as it was reported during the timeframe being examined, and does not rely on the ability of respondents to recall data. Kivijarvi and Saarinen (1995) successfully combined survey and secondary data from Finnish firms. They found a long-range correlation between investment in information services and maturity of information services and a correlation between maturity of information services and improved company performance. This study did not use the decision to outsource or degree of outsourcing as a dependent variable. In addition, a potential limitation is that research was confined to Finnish firms only. Grover, Cheon and Teng (1996) used the same dataset from their 1995 study to examine the issue of asset specificity and break down five functional areas for analysis:
Application development and maintenance
30
Systems operations Telecommunications End-user (help desk, desktop) Systems planning and management
Their approach shows, as other studies have, that the IT function can be segmented in a variety of ways. Satisfaction of the respondent (IT executives and managers) is used as the dependent variable. The survey obtained a 19% response rate and the significant coefficients suggested that generic or commodity-type outsourcing is more successful than those with high asset specificity, validating theoretical suppositions. The limitations of the respondents’ ability to recall data from three-years previous (see Teng, Cheon, Grover, 1995 above) is equally applicable to this study. Mitra and Chaya (1996) analyzed a dataset of 400 large and medium-sized US corporations obtained by mail and interview information collected by the editors of Computerworld and matched with accounting data obtained from Compustat. The results indicated that larger firms spend more on IT than smaller firms, possibly because of the higher control and monitoring costs of large firms. The results also indicated that IT might decrease these costs. The study found that IT investment results in lower average production costs but higher average overhead costs. Interestingly, there was no evidence of a reduction in labor costs. The study focused on the impact of IT on firm performance but did not measure possible determinants of outsourcing by those firms. In an examination of the U.S. and Japanese automobile industry Dyer (1997) finds, contrary to theoretical predictions, specifically by Williamson (1975), that high asset specificity may imply lower transaction costs. Dyer surveyed the two Japanese automakers (Nissan and Toyota), all three U.S. automakers, a sample of their suppliers,
31
and obtained 192 usable responses. The study identifies several potential variables that transaction cost theory may have omitted. The methodology used by Dyer may have also led to different empirical results because of his use of longitudinal data in the study, as opposed to cross-sectional data, the more common approach in the literature. As a proxy for transaction costs, Dyer used the dollar value of goods procured in 1991 divided by the number of people in the procurement function for the automakers. Since the data had to be obtained from a survey and not secondary sources, the indicator has the advantage of being specific, as well as being confined to the procurement area, which is generally well bounded most corporations. It may be however that procurement-related activities (which have transaction costs associated with them, such as much of the information gathering) occurred in other departments. If so, this would have had the effect of biasing downward the indicator for the number of people involved, along with the actual costs incurred in the procurement function. Dyer’s study suggests that for long-term economic relationships, search and transaction costs may be higher than predicted by Williamson. The author concludes that the Japanese appear to have lower transaction costs, even with high asset specificity, because of the following factors:
Small set of suppliers Economies of scale with regard to the small supplier group Extensive information sharing (less asymmetry of information) Greater degree of goodwill, trust Investments in co-specialized assets
Poppo and Zenger (1998) compare services and manufacturing companies with regard to determinants of boundary choice. The study used a blocking approach to separate firms that were primarily outsourced from those that performed most IT
32
functions internally. From their data, it is possible to infer some results regarding outsourcing even though it was not used as the dependent variable. It appears, for example, that services companies may be outsourcing to a greater degree. Further, in periods of rapid change in the information services environment, there may be a greater tendency to outsource. The authors admit that their methodology may have been affected by the rapid change that was occurring during the period during which the study was conducted. Specifically, the decision criteria for managers were unstable and not necessarily clear. The study indicated that greater asset specificity appears to lead to a greater degree of insourcing, as predicted by transaction cost theory. The authors analyzed exchange attributes of both the firm and the market in order to determine how make or buy decisions are made. The dataset used was a 1992 survey administered to top computer executives; 181 responses were received and 152 of the surveys were usable. Ang and Straub (1998) surveyed senior IT managers in 243 U.S. banks and found that perceived comparative advantages in production costs offered by outsourcers appeared to lead to greater outsourcing. Firm size was found inversely related to the degree of outsourcing by banks. Finally, high perceived transaction costs were also inversely related to the degree of outsourcing by banks. While the study does rely on survey feedback to obtain data on degree of outsourcing, it probed for responses from eight functional areas and used a 7-point Likert scale to identify the degree of outsourcing. The eight functional areas were:
Information Systems Strategy Information Technology Planning Capacity Management Production Scheduling Human Resources
33
Security Management Network Mangement Personal Computer (PC) Mangement
The study was, however, limited to the banking industry only and did not attempt to measure organizational focus as a determinant of outsourcing.
2.1.6 More Elaborate Models The last group of studies has focused on more econometric-style research and has probably had the fewest contributors to-date. This can be attributed to at least two factors. One is the relative infancy of outsourcing as a governance mechanism and the associated challenge, for many years, of satisfactorily framing the appropriate questions. The other factor is the difficulty in obtaining firm information based on SEC reporting requirements, as they do not lend themselves to the reporting of outsourcing-related information. Many of the variables that could potentially be used to measure outsourcing activity have actually been developed for measuring other aspects of the IT function. Brown, Gatian and Hicks (1995), for example, undertook a multiple-year (12 years) interrupted time-series analysis. A Strategic Information Systems (SIS) theory, which suggests that IT can produce improved business performance relative to other companies, was examined to try to find a correlation with financial performance. The event that triggered inclusion of the firm into the study was the year in which public knowledge became available that the firm had made significant investment in IT and was recognized as a leader. Measures of performance were then taken six years before and after the event. These measures were:
Growth - Year-to-Year percent in change of sales
34
Productivity - Sales per employee - Income per employee - Accounts receivable turnover - Inventory turnover - Asset turnover Profitability - Return on assets - Return on sales
Their study found significance for the SIS theory, particularly in the years following public awareness of the significant investments made by the firms to obtain competitive advantage. The sample consisted of 35 firms based on press reports of companies’ effective use of information systems. The final set of usable companies was limited because of three filters used to obtain the sample. First, the companies had to be identified as leaders in the use of IT, in the sense that that they were better than their peers. 4 Second, regulated companies were eliminated from the study, such as banks and insurance companies. Finally, companies that did not have financial information available on Compustat were also eliminated. While the study explored the performance of firms based on IT investment, outsourcing was not explicitly examined. Dewan, Michael and Min (1998), found that less vertical integration leads to higher levels of IT investment. In addition, related diversification requires greater IT investment than unrelated diversification. Firm diversification in general leads to greater IT investment. Finally, firms with fewer growth options have higher investments in IT. The importance of the study is that it highlights two potential measures of organizational focus (where reliable proxies have been elusive) that are objective and use refined scaling techniques. These measures were termed entropy measure of related/unrelated 4
These companies were cited in Wiseman, C. Strategic Information Systems. Homewood, IL: Irwin, 1988.
35
diversification and Vertical Industry Connection (VIC) index. The authors did not examine these measures in relation to outsourcing but did in fact suggest their use for further research in other areas, specifically mentioning outsourcing. Other articles which develop measures of IT performance, but which did not use outsourcing as the dependent variable and are outlined in Table 2.2 include Kettinger (1994), Mahmood and Mann (1993), Ou and Penman (1989), Palepu (1986), and Ohlson (1980). These studies do highlight the use of potential dependent variables that have been used to answer questions regarding how well IT has helped organizations perform. In one of the first statistical samples dealing with outsourcing, using mainly secondary sources, Loh and Venkatraman (1992) looked at several factors obtained from publicly reported firm data that might influence the decision to outsource. Results of the study found support for overall business cost structure and IT cost structure being positively correlated with degree of outsourcing. IT performance was negatively related to outsourcing. Hypotheses concerning financial leverage being positively related to outsourcing and business performance being negatively related to outsourcing were not supported. The study was largely limited to banks and other financial institutions, and hence its method of normalizing IT expenditure (dividing by assets) may not generalize well to non-banking firms. Finally, in one of the few other studies using secondary data combined with a more elaborate methodological analysis, and actually dealing with outsourcing, is a recent article by Smith, Sabyasachi, and Mitra (1998). The article builds on the work done by Loh and Venkatramen (1992) and identifies five themes that have emerged from as to why firms outsource:
36
1. 2. 3. 4. 5.
Cost reduction Focus on core competencies Liquidity needs IS capability Environmental factors
The study found significance only with regard to the need to reduce costs and generate cash (liquidity needs). Perhaps most interestingly, no significance was found with regard to variables associated with organizational focus (core competence). Because of the nature of the dependent variable however (only large-scale outsourcing agreements), the sample size was limited to 29 companies.
2.1.7 Opportunities for Extending Existing Work As evidenced above, studies on IT outsourcing have continued to extend research methodologies and improve understanding of outsourcing as a governance mechanism. Starting with conceptual frameworks and then moving from applied, to case study, the research has increasingly progressed to more rigorous and objective methodologies, the literature has followed a continuum of sorts. Initially, a body of research pursued the case study approach in order to map potential variables for future, more systematic research. As such, case studies have been helpful in providing an in-depth look at the nature of IT outsourcing and identifying potential variables for analysis. However, this early line of research may be reaching a point of diminishing returns. Increasingly work on outsourcing is becoming more systematic in nature, and is a logical extension of the conceptual and case study work that has gone before. Now researchers appear to be focusing on data obtained from wider, more representative
37
samples. Both primary and secondary data have been employed. The use of surveys, as an example of primary data collection, has been widely adopted because of the difficulty in otherwise obtaining data on outsourcing from secondary sources. While effective at allowing the researcher to tailor questions to their needs, they are nonetheless subject to bias on the part of the respondent. There is also the weakness of low response rates, although this concern appears to have been adequately addressed in the literature (Grover, Cheong, Teng 1996), i.e., low response rates do not necessarily imply bias in the sample. As an aside, a rare example of comparatively theoretical, almost purely economic approach to outsourcing bidding develops a mixed-integer programming model (Chadhury, Nam and Rao 1995), but does not offer any empirical evidence. The authors propose a precontract bidding mechanism for minimizing outsourcing costs, and indicate that one of the reason’s firms may outsource is to take advantage of the comparative scale economies that the outsourcer can bring to bear. The Dewan, Michael and Min (1998) study has potential applicability to outsourcing, which the authors in fact suggest as a research opportunity. As mentioned above, their study focused more generally on the IT function. But of significance is the fact that their approach has refined the methodology of examining organizational focus, perhaps better known as the concept of core competence, by using a methodology to measure related vs. unrelated diversification, which in turn derives primarily from several previous studies. Jaquemin and Berry (1979) developed an entropy measure of diversification for a study on corporate growth. Maddigan (1981) introduced a proposed measure for vertical integration, which unfortunately can only be computed periodically because of the need to obtain industry input-output data from the Bureau of Economic
38
Analysis. The most recent year for which such data is available is 1992. Palepu (1985), in a study of firm diversification and profitability developed index measures of degrees of related and unrelated diversification. Hill (1988) examined issues of organizational form, again with similar measures of related/unrelated diversification. Each of these studies used SIC codes to establish a reasonable proxy for organizational focus that is objective and readily available in longitudinal form through such sources as Compustat. Again, with regard to such studies that actually focus on outsourcing, Loh and Venkatraman (1992) operationalize the degree of outsourcing by using the ratio of IT outsourcing expenditure divided by total assets to normalize level of outsourcing by firm size. The study used both secondary data from publicly available sources and primary data from G2 Research Smith, Sabyasachi, and Mitra (1998) appear to have done the most with respect to outsourcing using only secondary data combined with rigorous methodologies and examining a variety of factors. The most significant limitation of the study were the small sample size, which required the use of non-parametric testing, and the narrow range of analysis that the dataset would support. Smith, et al. found support for the hypothesis that firms which outsourced were more cost conscious and than firms which did not. Also, the firms that outsourced IT tended to use the transaction as a means to generate cash. No evidence was found for the proposition that firms use of IT outsourcing as a means to focus on core competency, with the measures being revenue/employees and revenue/assets ratios. Similarly, lower profitability was not found to be significant in explaining why companies chose to outsource IT. While the study used a relatively small sample size of firms with major IT
39
outsourcing initiatives, the authors examined financial data from other firms in related industries as a form of control, and also as a way to adjust for industry-specific effects that may have impacted the decision to outsource. Since the dissertation will be using a larger sample, industry-effects can be isolated and examined through the use of dummy variables. Some of the difficulties associated with measuring outsourcing activity by firms have to do with the paucity of information and the elusiveness of satisfactory measurements. Firms do not regularly report their outsourcing activities or the dollar amounts spent on them. Further, IT expenditures tend to permeate the organization and often cannot be captured by a single departmental budget. Probably the easier approach to use when measuring outsourcing is a dichotomous measure, which is the method used by the studies listed in Table 2.1a. This has the advantage of relying exclusively on publicly available information. A more robust alternative, which invariably relies on surveys, is the use of a comparatively continuous measure of outsourcing activity which attempts to capture the degree of outsourcing by a particular firm using either a Likert scale or IT dollars spent (normalized in some fashion). Studies which use a dependent variable of this type are outlined in Table 2.1b. Independent variables and their roots in the literature are outlined in Table 2.2. These are discussed in more detail in Section III. This dissertation will attempt to fill the gap between the two types of studies indicated above. Specifically, a continuous measure of outsourcing is available from the INPUT database in the form of dollars spent on IT outsourcing. Further, a relatively large sample size (n=299) was derived from the nearly 2000 outsourcing incidents recorded, and these data are available over a period of roughly
40
ten years. This longitudinal picture of outsourcing makes more robust analyses possible, such as changes in the characteristics of how firms have chosen to employ outsourcing over the past decade, and the types of outsourcing in which they have chosen to engage.
41
Table 2.1.1a - Analysis of Dependent Variable Dependent Variable: Individual Outsourcing Agreements Author(s)
Indicator
Scope
Country
Unit of Analysis
Poppo and Zenger (1998) Smith, Mitra, Narasimhan (1998)
Dichotomous measure of nine functions least/most likely to be outsourced Announcements of large-scale outsourcing events
All industries
USA
152 firms; top computer executives
TimePeriod 1992
All industries
USA
29 incident firms
1985-1994
Dyer (1997)
Measures of asset specificity and transaction costs associated with outsourcing Outsourcing broken by two functional areas: operations (outsourced) and applications (not outsourced) Degree of outsourcing broken down by seven functional areas dichotomously (outsource: yes/no)
Auto industry
USA, Japan
192 firms
1996
Banking industry
USA
1 firm
1994
Cross sectional
All industries
USA
108 firms; senior IT managers from a sample obtained from the S&P Standard Register
1993
Cross sectional
Data Structure Cross sectional Longitudinal 3 Years (2 data points) Cross sectional Cross sectional
Palvia (1995)
Arnett and Jones (1994)
Data Structure Cross sectional Time series 3 Years, Annual Cross sectional
Table 2.1.1b - Analysis of Variables and Diagnostic Results Dependent Variable: Degree of Outsourcing by Firms Author(s)
Indicator
Scope
Country
Unit of Analysis
Ang and Straub (1998)
Banking industry
USA
243 US Banks
Teng, Cheon, Grover (1995)
Degree of outsourcing based on 7-point Likert scale response, broken down by eight functions Change in outsourcing budget as a percent of total IS budget
TimePeriod 1993
All industries
USA
188 firms; top computer executives
1992
Saarinen and Vepsallainen (1994) Loh and Venkatraman (1992)
Survey response indicating the degree of use of outside contractors (5-point scale) Ratio of IT outsourcing expenditure to total assets
All industries
Finland
55 Finnish firms
1988
Insurance industry
USA
57 Firms available through both G2 Research and Compustat II or Lotus’ CD/Corporate
1989
42
Table 2.1.2 - Analysis of Variables – Independent Variables Context Cash Needs
Independent Variable Liquid Assets on hand
Ability to service debt(1)
Ability to service debt(2)
Ability to service debt(3)
Financial leverage
Cost Focus
Overhead expenses
Operating expenses
Author(s) Smith, Mitra, Narasimhan (1998) Ang and Straub (1998) Palepu (1986) Loh and Venkatraman (1992) Smith, Mitra, Narasimhan (1998) Ohlson (1980) Smith, Mitra, Narasimhan (1998) Ohlson (1980) Smith, Mitra, Narasimhan (1998) Segars and Grover (1995) Kettinger (1994) Ohlson (1980) Smith, Mitra, Narasimhan (1998) Palvia (1995) Segars and Grover (1995) Kettinger (1994) Palepu (1986) Smith, Mitra, Narasimhan (1998) Mitra and Chaya (1996) Smith, Mitra, Narasimhan (1998) Mitra and Chaya (1996) Loh and Venkatraman (1992) Loh and Venkatraman (1992)
Operational Definition Cash and cash equivalents / Revenues
Findings -
Avg 3-yr retained earnings (for each bank) – Avg 3-yr retained earnings(all banks) Cash and cash equivalents / Total Assets Earnings per share Total liabilities / Revenues
n.s.
Total liabilities / Total Assets Long-term debt / Revenues
n.a. +
Long-term debt / Total Assets Current liabilities / Revenues
n.a. +
Current assets / Current liabilities Current assets / Current liabilities Current liabilities / Total Assets Total liabilities / Shareholder equity
n.a. n.a. n.a. +
Level of non-productive assets Shareholder equity / Debt Shareholder equity / Debt Total debt / Shareholder equity SG&A / Revenue
+ n.a. n.a. n.a. -
SG&A / Revenue (CGS+ SG&A) / Revenues
n.a. n.s.
(CGS+ SG&A) / Revenues (CGS+ SG&A) / Revenues (CGS+ SG&A) / Total assets
n.a. + +
n.a. n.s. n.s.
Significant relationships: positive relationships are indicated by ‘+’, negative relationships are indicated by ‘–‘, and non-significant relationships are indicated by ‘n.s.’. n.a. - dependent variable was other than outsourcing. Bold indicates independent variables which have actually been tested as possible determinants of outsourcing. Other variables are measures of firm performance which have been proposed in the literature and may be potential predictors of outsourcing. CGS-Cost of goods sold SG&A-Selling, general and administrative IBTDE- Income before taxes, depreciation and extraordinary items
43
Table 2.1.2 Analysis of Variables – Independent Variables (cont.) Context Company size
Organizational Focus
Independent Variable Revenues
Employee productivity
Asset productivity
Entropy measure of related/unrelated diversification
Vertical Industry Connection (VIC) index
Author Smith, Mitra, Narasimhan (1998) Ang and Straub (1998) Mitra and Chaya (1996) Chadhury, Nam, Rao (1995) Kettinger (1994) Smith, Mitra, Narasimhan (1998) Segars and Grover (1995) Brown, Gatian, Hicks (1995) Kettinger (1994) Smith, Mitra, Narasimhan (1998) Brown, Gatian, Hicks (1995) Kettinger (1994) Loh and Venkatraman (1992) Dewan, Michael, Minn (1998) Hill (1988) Palepu (1985) Jacquemin and Berry (1979) Dewan, Michael, Minn (1998) Maddigan (1981)
Operational Definition (abbreviations detailed below) Revenues adjusted for inflation
Findings n.s.
Log of firm size based on assets Revenues Scale of firm’s IT relative to outsourcer Gross value of assets Revenue / Employees
n.a n.a. n.s.
Revenue / Employees Revenue / Employees; Income / Employees Revenue / Employees Revenues / Assets
n.a. n.a. n.a. n.s.
Revenues / Assets; Revenues / Acct Rec; Revenues / Assets Revenues / Assets Revenues / Assets; Income / Assets Related diversification / Total diversification
n.a.
Related diversification; Unrelated diversification Diversification index Entropy measure of diversification and firm growth Index measure approaching 1 as degree of vertical integration increases Index of vertical integration
n.a. n.a. n.a. n.a.
n.a. n.s. n.a.
n.a.
Significant relationships: positive relationships are indicated by ‘+’, negative relationships are indicated by ‘–‘, and non-significant relationships are indicated by ‘n.s.’. n.a. - dependent variable was other than outsourcing. Bold indicates independent variables which have actually been tested as possible determinants of outsourcing. Other variables are measures of firm performance which have been proposed in the literature and may be potential predictors of outsourcing. CGS-Cost of goods sold SG&A-Selling, general and administrative IBTDE- Income before taxes, depreciation and extraordinary items
44
Table 2.1.2 Analysis of Variables – Independent Variables (cont.) Context Profitability
Independent Variable Return on assets or use of assets
Return on equity
Operating margin (operating profit)
Growth
Revenue growth
Author Smith, Mitra, Narasimhan (1998) Brown, Gatian, Hicks (1995) Mahmood and Mann (1993) Ou and Penman (1989) Smith, Mitra, Narasimhan (1998) Palepu (1986) Smith, Mitra, Narasimhan (1998) Mitra and Chaya (1996) Smith, Mitra, Narasimhan (1998) Brown, Gatian, Hicks (1995) Mahmood and Mann (1993)
Operational Definition (abbreviations detailed below) IBTDE / Assets
Findings n.s.
IBTDE / Assets IBTDE / Assets IBTDE / Assets IBTDE / Shareholder equity
n.a. n.a. n.a. n.s.
IBTDE / Shareholder equity 1 minus operating expense / Revenues
n.a. n.s.
1 minus operating expense / Revenues Yearly percentage change in sales
n.a. -
Yearly percentage change in sales Five-year revenue growth rates
n.a. n.a.
Significant relationships: positive relationships are indicated by ‘+’, negative relationships are indicated by ‘–‘, and non-significant relationships are indicated by ‘n.s.’. n.a. - dependent variable was other than outsourcing Bold indicates independent variables which have actually been tested as possible determinants of outsourcing. Other variables are measures of firm performance which have been proposed in the literature and may be potential predictors of outsourcing. CGS-Cost of goods sold SG&A-Selling, general and administrative IBTDE- Income before taxes, depreciation and extraordinary items
CHAPTER THREE RESEARCH DESIGN AND MODEL DEVELOPMENT The scholarly literature reviewed in the previous chapter identified a variety of influences on the decision to outsource. In other cases, influences of IT on firm performance are introduced, but not applied to the question of outsourcing. Because I wish to test, extend and clarify several of these relationships and organize a still emerging area of research, I begin this chapter by framing several research questions and then proceed to discuss key methodological issues. These include the overall conceptual framework of the study, the conceptualization of the dependent and predictor variables, measurement and data development issues, strategies for the control of unwanted influences, and issues related to the use of OLS regression analysis as a means to examine hypothesized causal relationships between variables.
3.1 Research Questions Reasons given for the marked increase in outsourcing, particularly over the past ten years appear to have shifted from economies to organizational focus, at least on an anecdotal basis. Organizational focus or core competency arguments are similar to resource-based theories discussed in economic literature, where the organization makes use of available resources and looks to the market to acquire what it cannot efficiently produce in-house. Why do firms choose to outsource? What objective determinants exist and can be measured in that regard. Further, once the decision to outsource has been made and
45
46
implemented, what is the impact on organizational performance. These are the primary questions that form the basis of this study.
3.2 Database Development and Concept Measurement In analyzing both determinants of outsourcing as well as performance results, the ability to operationalize data is inevitably limited to the available measures. While publicly reported data are somewhat inflexible in format and content, they also has the virtue of being comparatively consistent and objective across firms. Operationalization of the variables derives specifically from three previous papers (Dewan, Michael, Min, 1998; Smith, Sabyasachi and Mitra, 1998; Loh and Venkatraman, 1992).
3.2.1 Unit of Analysis The unit of analysis is a set of outsourcing events covering the 1990-1999 time period. The sample frame consists of a database obtained from INPUT, a web-based market research and marketing services firm which has tracked IT outsourcing events over the past ten years. Data from the INPUT sample was matched with SEC financial and SIC data obtained from the CompuStat database. The selection of the data were driven by theoretical propositions as to why firms outsource, and financial measures of firm performance and efficient use of inputs.
3.2.2 Firm Categorization All of the firms being studied were identified by Standard Industrial Classification
47
(SIC) codes. However, the SIC has become increasingly dated and has been replaced by the North American Industrial Classification System (NAICS), which is intended to better reflect the changing industry structure of the economy. Use of the SIC codes is somewhat unwieldy in some respects. While SIC codes provide a high level of detail, which can be useful for in-depth analysis, for general classification purposes such detail is not necessary or even helpful. Secondly, although SIC codes can provide a high degree of specificity with regard to activities such as manufacturing or banking, for other areas such as business services, many highly disparate industries are grouped together. For this reason, firms in the sample have been categorized using Information Week’s 20 industry classification which groups firms into the following categories:
Banking and Financial Services Professional Services Telecommunications Health Care Information Technology Media and Entertainment Consumer Goods Insurance Construction and Engineering Electronics Energy Pharmaceuticals and Medical Equipment Transportation Chemicals Manufacturing Retail and Distribution Utilities Hospitality and Travel Metals and Natural Resources Food and Beverage Processing
The above categorization should serve to divide companies into enough groups to provide meaningful detail, yet not be overly granular. While SIC codes were still used for some
48
of the analysis, this grouping was useful in isolating industry effects on determinant and performance measures in the regression analysis.
3.3 Determinants of Outsourcing The applicable causal diagram and justifications for the operationalized variables for outsourcing determinants are outlined in Figure 3.1. Y = f (X 1 , X 2 , X 3 , …, X 16 ) Figure 3.1.1 -- Causal Diagram
Profitability
Company Size
Revenue Growth Outsourcing Intensity Cost Focus
Cash Needs
Organizational Focus
49
Profitability X 1 = Return on assets X 2 = Return on equity X 3 = Operating margin Company size X 4 = Revenues Revenue Growth X 5 = Year-over-year change in revenues Cost focus X 6 = Overhead expenses X 7 = Total expenses
Cash needs X 8 = Liquid assets on hand X 9 = Ability to service debt (total liabilities divided by revenues) X 10 = Ability to service debt (long-term debt divided by revenues) X 11 = Ability to service debt (current liabilities divided by revenues) X 12 = Financial leverage
Organizational focus X 13 = Revenue divided by number of employees X 14 = Asset turnover X 15 = Entropy measure of related vs. unrelated diversification
3.3.1 Dependent Variable Each occurrence of outsourcing from the INPUT database is treated as an independent observation. The data span a ten-year period, from 1990 to 1999. INPUT tracks the annual dollar amount spent on outsourcing for a particular deal. However, in order for the events to be comparable, the magnitude of the outsourcing must be normalized against some measure of firm size. Loh and Venkatraman (1992) have computed operationalization of the variable “level of outsourcing” as follows:
Ratio of IT outsourcing expenditure to total assets
The Loh and Venkatraman study however used banks and financial institutions as the primary unit of analysis, so it would useful to have another method of normalizing the outsourcing expenditures. Since revenue is often used to normal data across firms, I propose a similar ratio for comparison purposes:
Ratio of IT outsourcing expenditure to total sales
In addition, a third dependent variable was operationalized:
50
Ratio of IT outsourcing expenditure to IT budget
While the INPUT database tracked outsourcing expenditures where the information was available, it did not track the IT budget of the firms. Information Week, however performs an annual survey and polls respondents with regard to their IT budget as a percent of sales. I used these percentages to estimate IT budgets as a point of comparison for normalization of degree of outsourcing, but because many of the independent variables make use of sales data to create ratios, the possibility of multicollinearity is present with this operationization of the dependent variable. As indicated above in the review of literature, categorization of IT has taken many forms and the boundaries have been established in many places.
Platform Operations Systems and Technology Services Application Operations Network/Carrier Management Business Operations Application Management Desktop Services Electronic Markets Management Consulting
As evidenced by the literature already cited, the definition of IT functions can vary. In order to be as clear as possible, INPUT’s criteria for categorizing outsourcing events are outlined below:
51
Table 3.3.1 – Outsourcing Function Definitions Functional Area
Description
Platform Operations
The vendor manages and operates the computer systems (typically mainframe or midrange) to perform the client responsibility for client. The vendor manages and operates the computer systems, to perform the client functions, without taking responsibility for the client systems.
Systems and Technology Services Application Operations
Combination of Platform Operations, systems integration and maintenance.
Network/Carrier Management
Business Operations
Application Management
Desktop Services
Electronic Markets
Management Consulting
The vendor manages and operates the computer systems to perform the client’s business functions, and is also responsible for maintaining, or developing and maintaining the client. Network Management: The vendor assumes responsibility for operating and managing the client. The client outsourcing contract may include only the management services or it may cover the full costs of the communications services and equipment plus the management services, plus voice/data convergence, IP telephony configuration and services. Business operations outsourcing (also known as business outsourcing or functional outsourcing) is a relationship in which one vendor is responsible for performing an entire business/operations function, including the information systems outsourcing that supports it. The information systems outsourcing content of such a contract must be at least 30% of the total annual expenditure in order for INPUT to include it in the outsourcing market. Examples of business operations that are outsourced include telephone company billing and employee benefits processing. The vendor has full responsibility for maintaining and upgrading some or all of the application systems that a client uses to support business operations and may also develop and implement new application systems for the client. An applications management contract differs from traditional software development in the form of the client/vendor relationship. Under traditional software development in the form of the client/vendor relationship. Under traditional software development services, the relationship is project based. Under applications management, it is time and function based. These services may be provided in combination or separately from platform outsourcing. The vendor (outsourcer) assumes reponsibility for the deployment, maintenance, and connectivity among the personal computers and/or workstations in the client organization. The services may also include performing the help-desk function. Equipment as well as services can be part of a desktop services outsourcing contract. Note: This type of client service can also be provided through traditional professional services where the contractual criteria of outsourcing are not present. The vendor takes responsibility for specialized e-business and/or e-commerce functions related to web-site design, hosting, management, security and maintenance. These functions may comprise either business-to-business, or business-to-consumer applications. Related functions such as e-mail management, web-enabled customer service, or set-up for future transaction processing. Due to problems with the definition of this term, it was subsequently dropped by INPUT. Initially, it was used to designate contracts that were characterized by a large component of management consulting as well as related IT outsourcing. IT strategy consulting was usually conceived as a preliminary to a more focused, clearly defined outsourcing contract. Because of the fact that this category was dropped during the period under examination, it was not included as a variable in the regressions which examine the type of function outsourced.
Literature References West, (1994) Grover, Cheon, Cheng (1996) Ang and Straub (1998) Grover, Cheon, Cheng (1996) West, (1994) Grover, Cheon, Cheng (1996) West, (1994) Grover, Cheon, Cheng (1996) Ang and Straub (1998) Ang and Straub (1998)
West, (1994) Grover, Cheon, Cheng (1996)
Grover, Cheon, Cheng (1996) Ang and Straub (1998)
Ang and Straub (1998)
52
3.3.2 Predictors and Covariates Since the data from CompuStat and INPUT are available in time-series, there is an opportunity to examine outsourcing over a ten-year timeframe. For the examination of the determinants of outsourcing, dummies have been included for the year of the outsourcing deal. For outsourcing determinants, data from Compustat was obtained for the three years prior to the outsourcing event. It seems unlikely that events prior to that time period would have an effect on the decision to outsource. The variables to be used in the model are outlined in Table 3.1.
3.3.2.1 Profitability H1: Firm profitability and outsourcing intensity are negatively related. It has been claimed that firms which outsource IT suffer from low profitability and performance and that the stock market reacts favorably to firm IT outsourcing announcements. Firms with low profitability may outsource IT for short-term cost reductions or to send a positive message to shareholders. Three measure of profitability that have been identified in the literature will be used here.
Return on assets or use of assets: income before taxes, depreciation, and extraordinary items divided by assets.
Return on equity: income before taxes, depreciation, and extraordinary items divided by common shareholder equity.
Operating margin or operating profit: 1 minus operating expense divided by revenues. Again, operating expense is measured as cost of goods sold plus sales, general and administrative expenses.
3.3.2.2 Company Size H2: Firm size and outsourcing intensity are positively related.
53
Because outsourcing intensity may be an artifact of company size, it may be worthwhile to examine organizational size for significance. This can probably best be proxied by revenues, adjusted for inflation. Again, this independent variable could be examined for its effect on Outsourcing Intensity in a longitudinal fashion In order to control for such effect, a measure of company size will be included in the regression.
Company size: real revenues (revenues adjusted for inflation)
3.3.2.3 Growth H3: Firm growth rates and outsourcing intensity are negatively related. Outsourcing may be a response to slower growth. Firms may shed assets and personnel in an attempt to become more productive or to obtain outside expertise.
Growth: yearly percentage change in sales
3.3.2.4 Cost Focus H4: Overhead expenses and outsourcing intensity are positively related. Because of agency costs, internal IT shops may not be run as efficiently as that of an outsourcer. Cost reduction and control are often given as reasons for outsourcing IT. Outsourcers are often believed to have better economies of scale, tighter control over fringe benefits, better access to low cost labor pools, and processes that are more efficient. Overhead expenses will be operationalized by two measures:
SG&A overhead expenses: a function of selling, general and administrative expenses divided by revenues.
Total overhead expenses: operating expense (cost of goods sold plus selling, general and administrative) divided by revenues.
54
3.3.2.5 Cash Flow H5: Cash flow and outsourcing intensity are negatively related. Firms often outsource IT to generate cash. The initial payment firms often receive when entering into outsourcing arrangements may be attractive to firms burdened with short-term liabilities and higher debt. Liquidity will be measured by:
Liquid assets on hand: cash and cash equivalents divided by revenues.
Three measures of a firms ability to service debt will be used:
Ability to service debt(1): total liabilities divided by revenues. Ability to service debt(2): long-term debt divided by revenues. Ability to service debt(3): current liabilities divided by revenues.
The degree to which a company relies on debt will be measured by:
Financial leverage: total liabilities divided by common shareholder equity.
3.3.2.6 Organizational Focus H6: Organizational focus and outsourcing intensity are positively related. Companies may outsource IT to simplify the management agenda and focus on the firm’s core business. This variable has limited treatment in the literature. The proxies used consist of the following:
Organizational focus(1): Revenues divided by the number of employees. Organizational focus(2): Asset turnover measured by revenues divided by assets.
In an attempt to develop a new measure of organizational focus, I apply one of the techniques used by Dewan, Michael and Min who proposed a measure of business diversification for measuring the boundaries of the firm with regard to IT functions and
55
might be instructive as a crude measure of corporate focus:
Organizational focus(3): Ratio of related lines of business to the total number of lines of business.
The Dewan et al. (1998) study examined the impact of diversification on the IT investment and suggested that the measure could be usefully applied examining the impact of diversification on IT outsourcing as well. It is important to note however that such a measure is not a reasonable proxy for core competence as defined by Prahalad and Hamel (1990). The two authors, now famous for coining the term explicitly argue that core competence can manifest itself by competing in a variety of seemingly unrelated industries, at least on the surface. Unfortunately this also makes their measure of core competence quite difficult to define. The measure of related vs. unrelated diversification has at least the virtue of being relatively objective. Measure of core competence in fact have often defied objective definitions and are often subject to either broad or vague definitions by consultants and managers. In addition to measures of related and unrelated diversification, Dewan, et. al.(1998) also used a measure of degree of vertical integration or Vertical Industry Connection (VIC) index. However this measure requires the use of an input-output matrix which is generated by the Bureau of Economic Analysis approximately every five years. This input-output matrix measures the relative level of interaction between industries (Maddigan 1981) and can be used to generate a sale-weighted overall index for each firm . However, because the data span ten years, an input-output matrix would be required for each year during the 1990 – 1999 timeframe in order to compute the VIC, but only 1992 is available.
56
3.3.2.7 Industry Effects With the use of dummy variables to analyze firms by industry, it is possible to isolate the effects of industry on the decision to outsource. Further, the use of dummy variable enables interaction terms to be generated and thus analyze how experience, for example may impact degree of outsourcing (or some other covariate). This method of analysis is commonly applied in the social science literature, but has not been widely used on research dealing with outsourcing impacts. There are no a priori indications that certain types of industries outsource more than others, or that types of functions outsourced has changed systematically over time, however the available data lends itself to this type of analysis and therefore suggests that the relationships could be profitably explored:
Banking and Financial Services Professional Services Telecommunications Health Care Information Technology Media and Entertainment Consumer Goods Insurance Electronics Energy
Pharmaceuticals and Medical Equipment Transportation Chemicals Manufacturing Retail and Distribution Utilities Hospitality and Travel Metals and Natural Resources Food and Beverage Processing Construction and Engineering
Of the companies which could be identified from the INPUT database for the determinants and effects samples, none were construction and engineering related (see appendix).
3.3.3 Summary of Variables The variables to be examined for determinants of outsourcing can be summed up
57
in Table 3.3.1.
58
Table 3.3.1 – Independent and Dependent Variables Dependent Variable: Code Indicator OI Degree of outsourcing by a firm
Independent Variables: Outsourcing Variable Name Determinants 1. Profitability
Operational Definition Ratio of IT outsourcing expenditure to total assets, with categorization by functional area
Predicted Relationship
Data Source INPUT Database, augmented by a fax survey and SEC documents
Variable Code
Operational Definition
Data Source
Return on assets or use of assets
-
ROA
Income before taxes, depreciation and extraordinary items divided by assets
Compustat
Return on equity
-
ROE
Income before taxes, depreciation and extraordinary items divided by common shareholder equity
Compustat
Operating margin (operating profit)
-
OPMG
1 minus operating expense divided by revenues
Compustat
2. Company Size
Revenues
-
REV
Revenues adjusted for inflation
Compustat
3. Growth
Revenue growth
-
REVGR
Yearly percentage change in sales
Compustat
4. Cost Focus
Overhead expenses
-
OH1
Sales, general and administrative expenses (SG&A) divided by revenues
Compustat
Total expenses
-
OH2
Cost of goods sold plus SG&A divided by revenues
Compustat
Liquid assets on hand
-
LQ
Cash and cash equivalents divided by revenues
Compustat
Ability to service debt
-
TLR
Total liabilities divided by revenues
Compustat
Ability to service debt
-
LDR
Long-term debt divided by revenues
Compustat
Ability to service debt
-
CLR
Current liabilities divided by revenues
Compustat
Financial leverage
+
LEV
Total liabilities divided by shareholder equity
Compustat
Predisposition to outsource (Degree of virtual organization)
+
REVEM
Revenues divided by number of employees
Compustat
Asset turnover
+
REVAT
Revenues divided by assets
Compustat
Entropy measure of related/unrelated diversification
+
RD
Related diversification divided by total diversification
2, 4 digit SIC codes (PSIC, SICNM)
5. Cash Needs
6. Organizational Focus
59
Table 3.3.2 – Correlation Matrix - Means Independent Variables (1) ROA (2) ROE (3) OPMG (4) REV (5) REVGR (6) OH (7) TLOH (8) LQ (9) TLR (10) LDR (11) CLR (12) LEV (13) REVEM (14) REVAT (15) RD n
(1)
(2)
(3)
1.00 -0.01 0.82 -0.02 -0.35 -0.80 -0.82 -0.81 -0.82 -0.82 -0.82 -0.01 0.02 -0.01 0.05
1.00 -0.01 -0.01 0.01 0.03 0.01 0.01 0.04 0.03 0.01 0.99 0.00 -0.01 -0.04
1.00 0.02 -0.39 -0.98 -0.99 -0.99 -0.97 -0.99 -0.97 0.02 0.03 0.02 0.05
297
296
298
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
1.00 -0.04 -0.02 -0.02 -0.01 -0.01 -0.01 -0.02 0.00 0.93 0.99 0.05
1.00 0.40 0.39 0.40 0.42 0.42 0.35 0.00 -0.05 -0.03 -0.23
1.00 0.98 0.99 0.97 0.99 0.97 0.01 -0.03 -0.02 -0.05
1.00 0.99 0.97 0.99 0.97 -0.02 -0.03 -0.02 -0.05
1.00 0.97 0.99 0.98 -0.01 -0.02 -0.01 -0.05
1.00 0.99 0.96 0.02 -0.02 -0.02 -0.06
1.00 0.97 0.01 -0.01 -0.01 -0.06
1.00 -0.01 -0.02 -0.02 -0.01
1.00 0.00 -0.01 -0.05
1.00 0.95 0.01
1.00 0.04
1.00
299
293
220
298
297
299
298
230
298
280
299
275
Table 3.3.3 – Correlation Matrix – Lagged 3 Years Independent Variables
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(1) ROA –3 (2) ROE –3 (3) OPMG –3 (4) REV –3 (5) REVGR -3 (6) OH –3 (7) TLOH –3 (8) LQ –3 (9) TLR –3 (10) LDR –3 (11) CLR –3 (12) LEV –3 (13) RVEM –3 (14) REVAT –3 (15) RD -3 n
1.00 0.02 0.43 0.10 -0.02 0.01 -0.43 -0.26 -0.33 -0.21 -0.38 -0.07 0.15 0.17 0.02
1.00 0.06 0.06 -0.04 -0.16 -0.06 -0.03 -0.01 -0.03 0.07 0.98 0.02 0.01 0.01
1.00 0.48 0.14 0.04 -1.00 -0.06 0.09 0.11 -0.01 0.02 0.64 0.57 0.00
1.00 -0.12 -0.17 -0.48 -0.09 -0.11 -0.11 -0.02 0.06 0.80 0.78 0.01
1.00 0.00 -0.13 0.24 0.18 0.34 -0.22 -0.08 0.08 0.05 -0.15
1.00 -0.04 0.29 0.26 0.25 0.45 -0.19 -0.18 -0.17 0.03
1.00 0.06 -0.09 -0.10 0.01 -0.02 -0.64 -0.57 -0.00
1.00 0.78 0.80 0.27 -0.15 -0.06 -0.07 -0.08
1.00 0.93 0.40 -0.11 -0.08 -0.13 -0.02
1.00 0.21 -0.16 -0.03 -0.07 -0.07
1.00 0.10 -0.25 -0.28 0.12
275
275
277
280
275
200
278
275
280
264
213
(12)
(13)
(14)
(15)
1.00 0.02 0.00 0.01
1.00 0.98 -0.06
1.00 -0.07
1.00
280
263
280
241
60
Table 3.3.4 – Correlation Matrix – Lagged 2 Years Independent Variables
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(1) ROA –2 (2) ROE –2 (3) OPMG –2 (4) REV –2 (5) REVGR –2 (6) OH –2 (7) TLOH –2 (8) LQ –2 (9) TLR –2 (10) LDR –2 (11) CLR –2 (12) LEV –2 (13) RVEM –2 (14) REVAT –2 (15) RD –2 n
1.00 0.01 0.40 -0.04 -0.00 0.01 -0.39 -0.13 -0.28 -0.23 -0.17 -0.02 0.01 -0.04 0.08
1.00 0.12 -0.01 0.04 0.16 -0.12 -0.05 0.15 0.20 -0.04 1.00 -0.01 -0.01 -0.05
1.00 0.61 0.12 0.05 -1.00 -0.03 -0.04 0.03 -0.00 0.11 0.61 0.61 0.08
1.00 -0.02 -0.14 -0.62 -0.07 -0.08 -0.04 -0.23 -0.01 0.98 1.00 0.04
1.00 0.02 -0.13 -0.09 0.17 0.25 -0.10 0.05 -0.05 -0.02 -0.14
1.00 -0.05 0.35 0.32 0.33 0.40 0.16 -0.16 -0.14 -0.04
1.00 0.03 0.04 -0.02 -0.00 -0.11 -0.61 -0.61 -0.08
1.00 0.26 0.22 0.25 -0.05 -0.07 -0.07 -0.11
1.00 0.96 0.26 0.16 -0.09 -0.08 -0.03
1.00 0.10 0.21 -0.05 -0.05 -0.08
1.00 -0.04 -0.26 -0.24 0.12
1.00 -0.01 -0.01 -0.05
1.00 0.98 0.02
1.00 0.03
1.00
289
288
291
293
281
208
291
288
293
271
223
293
269
293
266
Table 3.3.5 – Correlation Matrix – Lagged 1 Year Independent Variables
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(1) ROA –1 (2) ROE –1 (3) OPMG –1 (4) REV –1 (5) REVGR –1 (6) OH –1 (7) TLOH –1 (8) LQ –1 (9) TLR –1 (10) LDR –1 (11) CLR –1 (12) LEV –1 (13) RVEM –1 (14) REVAT –1 (15) RD –1 n
1.00 0.09 0.37 0.11 -0.15 -0.35 -0.37 -0.36 -0.38 -0.37 -0.41 0.02 0.19 0.14 0.11
1.00 0.10 -0.01 -0.05 -0.10 -0.10 -0.10 -0.10 -0.10 -0.09 0.75 0.03 0.12 0.01
1.00 0.07 -0.43 -0.99 -1.00 -1.00 -0.99 -1.00 -0.95 0.02 0.01 0.11 0.06
1.00 -0.13 -0.08 -0.07 -0.07 -0.08 -0.08 -0.04 -0.03 -0.03 -0.02 0.19
1.00 0.44 0.43 0.43 0.43 0.43 0.36 -0.09 -0.20 -0.04 -0.15
1.00 0.99 0.99 0.99 0.99 0.95 -0.03 -0.03 -0.15 -0.06
1.00 1.00 0.99 1.00 0.95 -0.02 -0.01 -0.11 -0.06
1.00 1.00 1.00 0.96 -0.03 -0.01 -0.13 -0.06
1.00 1.00 0.97 0.00 -0.02 -0.17 -0.05
1.00 0.96 -0.01 -0.01 -0.16 -0.06
1.00 0.04 -0.05 -0.25 0.00
1.00 -0.02 -0.09 0.00
1.00 0.08 -0.06
1.00 -0.16
1.00
289
288
287
294
289
208
284
290
292
275
222
291
266
291
270
Several variables exhibit high degrees of mulitcollinearity, particularly for the mean independent variables.
(15)
61
3.3.4 Outsourcing Determinants The first regression estimation then is provided as follows: OI = β 0 + β 1 OH1 + β 2 OH2 + β 3 LQ + β 4 TLR + β 5 LDR + β 6 CLR + β 7 LEV + β 8 REVEM + β 9 REVAT + β 10 ROA + β 11 ROE + β 12 OPMG + β 13 DIV + β 14 REVGR + β 15 REV + µ i Where:
Outsourcing intensity (three measures) OIREV = Outsourcing outlay divided by revenues) OIAT = Outsourcing outlay divided by assets OIIT = Outsourcing outlay divided by estimated IT budget OH = SG&A divided by revenues TLOH = CGS plus SG&A divided by total revenue LQ = Cash and equivalents divided by revenues TLR = Total liabilities divided by revenues LDR = Long-term debt divided by revenues CLR = Current liabilities divided by revenues LEV = Total liabilities divided by common shareholder equity REVEM = Revenue divided by number of employees REVAT = Revenue divided by assets ROA = Income before taxes, depreciation, and extraordinary items divided by assets ROE = Income before taxes, depreciation, and extraordinary items divided by common shareholder equity
62
OPMG = 1 minus operating expense divided by revenues. Again, operating expense is measured as cost of goods sold plus sales, general and administrative expenses DIV = Entropy measure of related vs. unrelated diversification REVGR = Annual growth in revenue REV = Company size adjusted for inflation
The data have been analyzed from two perspectives. The larger dataset (n=299), where the unit of analysis is dollar amounts associated with outsourcing contracts, is used to analyze determinants of outsourcing. Examining the results of outsourcing is a more difficult undertaking because of the relative paucity of data (n=124). Nonetheless, even from the smaller dataset some preliminary conclusions can be drawn.
3.4 Research Design and Data Development Issues One of the key challenges in performing research on outsourcing, as with many other aspects of organizational performance and governance, is access to data and the means by which it is obtained. As mentioned previously, studies on outsourcing have tended to rely on relatively short-term surveys and case studies. Surveys enable the researcher to tailor questions to the specific topic under investigation and, low response rates notwithstanding, are considered a reliable method in the literature. Surveys however have limited use in obtaining longitudinal data and are subject to bias on the part of respondents. Another popular method of obtaining data is by means of the case study. Case studies can provide a depth of understanding that is simply not feasible in other venues,
63
and can be very useful in framing issues associated with outsourcing. However, because firms must agree to participate, there are often issues of selection bias which may jeopardize representativeness of the sample. Even with studies using public data, researchers have had to settle for relatively small sample sizes. Because the SEC has no reporting requirements with regard to outsourcing, obtaining data for data for analysis has been difficult. INPUT, a research firm, has been tracking outsourcing events by firms for a number of years and undertakes a systematic process to identify and include these events into its database. While several research organizations track and project aggregate IT outsourcing trends such as International Data Corporation (IDC) and Gartner Group, these organizations do not maintain databases of outsourcing events in the detail that INPUT does, and report only the 100 largest outsourcing contracts in any given year. Data attributes tracked by INPUT include the date of the outsourcing event (announcement date), vendor, customer, type of function outsourced, components included, contract term, total contract value, industry, primary country affected by the outsourcing services, and in many cases a brief description of the nature of the deal. In some cases, the data was not available for each year. Segment data and associated revenues used to calculate diversification was not available prior to 1992.
3.4.1 Quality of the Data INPUT collects its data on outsourcing events by performing daily Internet sweeps and consulting public news sources such as the Wall Street Journal and the Financial Times. It consults non-U.S. newspapers in countries such as Australia, New Zealand,
64
Japan, as well as Europe. INPUT also monitors company press releases regarding outsourcing and consults vendor websites. In addition, INPUT frequently contacts participants in outsourcing events in order to procure additional data and clarify details on information obtained from other sources.
3.5 INPUT Database – Summary Statistics The INPUT Database consists of attributes of nearly 2,000 outsourcing contracts spanning almost a decade (1990-1999). Further, work on the database is ongoing. A total 1979 outsourcing events were available for analysis. Of this total, 557 events were IT outsourcing events pertaining to non-firm organizations such as government agencies at various levels and associations. The non-firm entities do not report the types of financial data required by the SEC for publicly-held companies, and thus were not suitable for analysis using the methodologies described earlier.
Table 3.5.1 – Firms vs. Non-Firms Firms Non-Firm Entities Total
Number 1422 557 1979
Percent 71.9 28.1 100%
The largest portion of non-firm entities was at the national or federal level (71%). The numbers fall off rapidly at the state and local level, and are even smaller for associations. Because of the lack of performance data on government entities, driven in part due to differing goals from the private sector in at least some respects, it was not possible to analyze them for this study.
Table 3.5.2 – Non-Firm Categorization by Type of Entity Classification (Non-FirmsEntities)
Number
Percent
65
National Governments and Agencies State Government and Agencies City/County Government and Agencies Associations Totals
397 83 58 19 557
71.3 14.9 10.4 3.4 100%
Outsourced functions for the organizations in the INPUT database fell into ten categories. The categorization was as follows: Table 3.5.3 – Categorization of IT Functions Outsourced – All Entities Function Outsourced (Entire INPUT DB) Systems and Technology Services Application Operations Business Operations Network/Carrier Management Application Management Desktop Services Platform Operations Electronic Markets Management Consulting Totals
# of Events 680 333 323 208 138 129 80 75 13 1979
% of Events 34.4 16.8 16.3 10.5 7.0 6.5 4.0 3.8 0.7 100%
For the firms (excluding government agencies, non-profit organizations and the like) in the INPUT database the categorization of outsourced functions was:
Table 3.5.4 – Categorization of IT Functions Outsourced: Firms Function Outsourced (Firms) Systems and Technology Services Application Operations Business Operations Network/Carrier Management Application Management Desktop Services Platform Operations Electronic Markets Management Consulting Totals
# of Events 380 272 252 168 110 100 70 61 9 1422
% of Events 26.7 19.1 17.7 11.8 7.7 7.0 4.9 4.3 0.6 100%
3.6 IT Outsourcing Determinants While the number of potential units of analysis from the INPUT database was
66
relatively large (n = 1422), the number of entries that could be definitively identified, were tracked by Compustat, and which had estimates of annual outsourcing available was much smaller. After eliminating those firms that met the above criteria, the dataset of outsourcing determinants was reduced to n = 299.
3.6.1 Summary Statistics – Determinants Dataset Given the difference between the size of the dataset from INPUT and the subsequent usable sample, it is appropriate to ask whether the sample population is representative. Table 4.2.1.1 shows a categorization by SIC category of each dataset, which suggests at least at face value, that no substantial differences exist between the two populations.
Table 3.6.1 – Categorization by SIC Code SIC Category
SIC Description
0000 1000 2000
Agriculture, Forestry, and Fisheries Mineral and Construction Industries Manufacturing (Food, Apparel, Lumber, Chemicals, Petroleum) Manufacturing (Rubber, Leather, Stone, Metals, Industrial, Electrical) Transportation, Communication and Utilities Wholesale and Retail Trade Finanace, Insurance and Real Estate Service Industries (Hotels, Business, Personal, Motion Pictures, Amusements) Service Industries (Health, Legal, Educational, Social, Museums, Engineering) Not Classifiable
3000 4000 5000 6000 7000 8000
INPUT Determinants Database Sample Share Share (%) ( %) 0.0 0.0 1.0 2.0 11.5 14.8 17.5
24.6
21.1 6.6 27.2 10.5
21.6 4.4 21.2 9.4
4.5
2.0
0.0
0.0
(n = 1422)
(n = 299)
As another point of comparison between the INPUT database and the determinants sample, the types of functions outsourced were compared. The determinants sample has a
67
significantly larger number of outsourcing contracts that fall into the category of systems and technology services, with the difference being spread relatively evenly across the INPUT database. For this comparison, there is no apparent a priori reason to expect this difference to impact the regression analysis. Table 3.6.2 – Categorization of Technology Functions Outsourced Function Outsourced
INPUT
Systems and Technology Services Application Operations Network/Carrier Management Business Operations Application Management Desktop Services Platform Operations Electronic Markets Management Consulting Totals
26.7 19.1 11.8 17.7 7.7 7.0 4.9 4.3 1.0 100%
Determinants Sample 40.5 14.7 9.0 19.7 3.7 4.3 2.7 4.7 1.0 100%
Figure 3.6.1 – Determinants Sample of Annual IT Outsourcing by Year Annual Number of IT Outsourcing Deals Dataset: Outsourcing Determinants (n=299)
100 80 60 40 20 0 1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
Table 3.6.3 – Determinants Sample of Annual IT Outsourcing by Year No. of Events % of Events
1990 3
1991 9
1992 7
1993 5
1994 13
1995 20
1996 32
1997 40
1998 78
1.0
3.0
2.3
1.7
4.3
6.7
10.7
13.4
26.1
1999 92 30.8
Totals 299 100%
68
The large number of outsourcing events later in the decade available from the sample appears to be consistent with the general trend toward an increase in outsourcing since the time of the Kodak outsourcing deal with IBM in 1989. While the data on actual outsourcing is not available in any systematic fashion during the last ten years, it is possible to piece together a modicum of evidence based on proprietary research by various firms which is outlined in Table 5.3.1.2. The research available in the form of estimates and forecasts on both IT outsourcing, as well as outsourcing of all functions, both in the US and worldwide, does suggest a general trend of the increased use of outsourcing as a firm governance mechanism. In 1990, the global IT market was $9B according to Templeton Research (Lacity and Willcocks, 2000). Based on the data compiled in Table 5.3.1.2 below, that number has increased significantly. By 1998, for example, the same forecasters estimate that global IT outsourcing had grown to $99B. Table 3.6.4 – Outsourcing Trend and Forecast Statistics by Research Firms 1995 IT Outsourcing (US) Outsourcing Institute International Data Corporation (IDC) Gartner/Dataquest IT Outsourcing (Worldwide) International Data Corporation (IDC) Templeton Research
1996
1997
1998
1999
2000
2001
2002
2003
42 -
17.3
82 -
-
-
184 -
33
-
-
26.3
30.3
34.8
39.9
45.3
51.6
-
-
-
36.7
-
-
-
-
58.6
-
-
99
-
-
-
120
-
164
-
-
319
-
-
141 136
-
-
-
-
-
-
-
121
-
-
-
235 286
-
-
-
-
633
(1990 estimate is $9B)
Outsourcing – All Functions (US) 100 Outsourcing Institute Outsourcing – All Functions (North America) 116 Dun & Bradstreet Gartner Group Outsourcing – All Functions (Worldwide) 76 International Data Corporation (IDC) 150 185 Dun & Bradstreet Gartner Group
All figures in $US billions Shaded figures indicate forecast numbers
69
Table 3.6.5 – Growth of Outsourcing Firms Over the Past 20 Years U.S. computer-services establishments Market value of publicly-traded U.S. computer-services companies
1980 26,370 $91B
1990 78,788 $106B
1999 211,323 $416B
Source: Federal Reserve Bank of Dallas 1999 Annual Report
CHAPTER FOUR OUTSOURCING AND FIRM GOVERNANCE
4.1 Results of Regressions – Determinants of IT Outsourcing Data for the independent variables were collected and analyzed for the three years prior to the outsourcing event. Data were analyzed by lagging the independent variables 1, 2 and 3 years prior to the outsourcing event. In addition, the mean of the three years prior to the outsourcing event was used as a predictor variable. It should be noted from the start that, during the 1990s, most firms probably outsourced some functions, many of those IT functions. Analysis of the dataset then clearly does not presuppose that the dependent variable is a measure of first-time outsourcing by a sample firm. Rather the independent variable is a measure of the degree of outsourcing undertaken by a firm, which may in fact have been preceded by previous greater or lesser outsourcing activity. At the same time, outsourcing activity has continued to grow significantly during the 1990s, from an estimated $9 billion in 1990 to $99 in 1998 (although estimates vary widely), which suggests that many of the events from the INPUT database represent either first-time outsourcing events or expanded scope by companies that were already engaged in outsourcing activity. Based on the correlation matrices (Tables 3.2, 3.3, 3.4, 3.5), several highly correlated variables were identified. Fortunately, there are multiple measures for the same concept in almost every instance, so eliminating highly correlated variables does not preclude analysis of the hypotheses. While the first set of regressions included all variables, for the second set of regressions, OMPG (operating margin), TLOH (sales,
70
71
general and administrative plus cost of goods sold), LQ (liquidity), TLR (total liabilities divided by revenues), LDR (long-term debt divided by revenues), and RA (revenues divided by assets) were eliminated as all were highly correlated (R > .90). This reduced the mean variance inflation factor (VIF) to less than 10 and improved the significance of the remaining independent variables. The main conclusions remained essentially unchanged. For the following hypotheses, data for the independent variables were lagged and regressed on the dependent variables for each of the three years prior to the outsourcing event. The dependent variables consist of OIIT (outsourcing expenditures divided by estimated annual IT budget), OIREV (outsourcing expenditures divided by annual sales), and OIAT (outsourcing expenditures divided by total assets) and are reported in three separate tables. The independent variables have been treated in two ways: the data for the independent variables was used for each of three years leading up to the outsourcing event as well as using the arithmetic mean of the independent variables for the three years 5 prior to the outsourcing event. For both sets of regressions, the 19 available industry groupings are introduced in order to control for industry effects. In general, when evaluating the results of the regressions, more emphasis is given to predictors or determinants of outsourcing in the 1 – 2 years prior to the outsourcing event, as opposed the 3 years prior. Particularly if there is conflicting coefficient signs between years, those closest to the outsourcing event would seem most likely to be predictors of the decision to do so.
5
In some cases, three years of data were not available. This occurs across all of the independent variables for the entire timeframe of the study. The mean for the years of available data was used in these cases.
72
H1: Firm profitability and outsourcing intensity are negatively related. For Hypothesis 1, there are three measures of profitability (return on assets, return on equity, operating margin). Operating margins was dropped from the regression because it was highly correlated with several other independent variables. The remaining independent variables were regressed on the three normalized values for outsourcing intensity. There are several significant coefficients across the regressions in Tables 4.1.1, 4.1.2, and 4.1.3, but results are somewhat mixed. Return on assets is negatively related with outsourcing intensity, while return on equity is positively related to outsourcing intensity. There is no readily apparent reason why firms which have strong returns on equity would outsource to a greater degree than others, or why firms with poor return on assets would outsource to a lesser degree than others. On this basis, support for H1 is inconclusive. Regressions run on each variable individually reinforce the effect, and as such, measures of profitability as predictors of outsourcing activity remain inconclusive.
H2: Firm size and outsourcing intensity are positively related. While it is at least intuitively appealing to presume that larger firms might outsource to a greater degree than smaller firms, if anything, the evidence suggests a slight effect to the contrary. It is important to note however that the dataset used (intersects between the INPUT and CompuStat databases) is composed exclusively of publicly-held firms, and correspondingly excludes almost by definition the smallest firms, many of which are in a pre-IPO (initial public offering) phase and still privately held. However within the parameters of the dataset (firms with 1999 real revenues between $1.4M and $173B), there is at least modest evidence a size effect for any of the three
73
measures of outsourcing intensity. While several of the coefficients are significant, the actual values are small. There appears to be modest evidence that firm size is in fact negatively related to outsourcing intensity, the opposite result expected for H2.
H3: Firm growth rates and outsourcing intensity are negatively related. The results suggest that somewhat faster growth lagged two and three years prior to the outsourcing event increase the degree of outsourcing by firms. This suggests that faster than average growth may incline firm management to look at outsourcing as a way to better accommodate growth. Firms may outsource IT because internal or organic growth of an IT infrastructure cannot be achieved quickly enough. Again, the results are somewhat counterintuitive, although it is certainly possible that outsourcing is increasingly being treated as a kind of joint venture, which firms often enter into to augment internal resources. In essence, it appears that firms with higher revenue growth rates have higher outsourcing intensity. This result supports comparatively recent work which suggests that younger firms and start-ups use outsourcing as a means to create a virtual organization, where literally all non-core functions are contracted out in order to bring a product or service to market more quickly.
H4: Firm overhead expense and outsourcing intensity are positively related. Because TLOH (total overhead: sales, general and administrative [SG&A] plus cost of goods sold [CGS] divided by revenue) and OH (SG&A divided by revenue) are highly negatively correlated (r = -0.98), TLOH was dropped from the regressions in order to reduce VIF resulting from multicollinearity. Here the results are resoundingly clear. Firms that have high overhead in turn have high outsourcing intensity across the board for
74
all lagged variables and the mean of the lags. The results strongly suggest that firms outsource as a way to gain control over general overhead costs.
H5: Cash flow and outsourcing intensity are negatively related. Several of the measures of cash flow are highly correlated (LQ [liquidity], TLR [total liabilities divided by revenues], LDR [long-term debt divided by revenues]) and have been removed from the regressions. This still leaves two measures of cash flow, CLR (current liabilities divided by revenue) and LEV (total liabilities divided by common shareholder equity). There appears to be strong support for Hypothesis 5. CLR is negative and significant when lagged two and three years prior to the outsourcing event. And LEV is negative and significant when lagged one year prior to the outsourcing event. Reducing the number of variables measuring cash flow from five to two, as mentioned earlier, greatly decreases the mean VIF and correspondingly increases the significance of the remaining two independent variables. Essentially, it appears that firms may outsource IT functions in response to cash flow problems.
H6: Organizational focus and outsourcing intensity are positively related. The last hypothesis deals with the assumption that organizations outsource because they want to focus on core competencies. As such, this would suggest that firms which have high outsoucing intensity have fewer employees involved in non-core activities and thus should have higher revenue per employee ratios (RE) or higher revenue to asset ratios (RA). RE and RA are highly correlated. Because RA is also highly correlated with other independent variables, it was dropped from the reported regressions. However, RA was substituted in a separate regression run; in neither case did
75
RE or RA have significant coefficients. The third independent variable used to measure organizational focus is the measure of related diversification (RD), which uses two and four digit SIC codes as developed by Maddigan (1981). The results of the regressions show that RD is actually negatively related to outsourcing intensity, which again, is somewhat counterintuitive. In effect, firms which tend to diversify narrowly appears to have a lower outsourcing intensity. It may be that firms with a narrow product focus are more progressive in terms of their IT capabilities, and may be better able to perform such functions than outsourcers or there may simply be a lack of sufficient outsourcers in the marketplace. Related diversification measures only the most salient aspects of a firm’s output, and has been largely biased toward manufactured goods over services. Firm focus, as measured by core competency is devilishly difficult to define. A firm that transports natural gas and owns the rights of way may decide that its core competency enables it lay fiber optic cable over those rights of way and sell broadband telecommunication services. Needless to say, the SIC codes for two such lines of business quite removed from one another. In reality, firms can be sliced, diced and categorized in a number of ways, many of which are not formally measured by government agencies or research firms. The significance of the RD coefficients then may lead to more questions than it answers, and objective measures of firm focus or core competence remain elusive.
76
Table 4.1.1 – Dependent Variable OIIT - Determinants of IT Outsoucing Intercept Return on Assets Return on Equity Real Revenue Revenue Growth Overhead Current Liabilities/Revenue Leverage Revenue/Employee Related Diversification Banking and Finance Professional Services Telecommunications Healthcare
0.520*** (0.170) -1.073** (0.543) -0.035 (0.055) -0.000 (0.000) 0.002** (0.001) 0.837** (0.342) -0.961** (0.435) 0.009 (0.012) 0.000 (0.000) -0.096 (0.115)
Information Technology Media and Entertainment Insurance Electronics Energy Pharmaceuticals and Medical Equip Transportation Chemicals Manufacturing Retail and Distribution Utilities Hospitality and Travel Mining and Natural Resources Food and Beverage Consumer Goods R2 Adjusted R2 N F-value Df
.1559 .0904 126 2.38 116
Lag 3 -0.453 (-0.453) 0.013 (0.013) -0.000 (0.000) 0.001 (0.001) 0.888** (0.888) -0.940** (-0.940) -0.002 (-0.002) 0.000 (0.000) -0.196 (0.139) 0.285 (0.459) -0.087 (0.420) 0.164 (0.411) 0.162 (0.504) 0.002 (0.413) -0.231 (0.431) -0.120 (0.563) 0.373 (0.411) -0.020 (0.438) -0.158 (0.452) 2.335*** (0.661) 0.164 (0.410) 0.213 (0.405) -0.156 (0.449) -0.044 (0.564) 0.128 (0.429) 0.795 (0.488) 0.330 (0.464) .4467 .3014 126 3.07 99
0.471** (0.165) -0.472 (0.734) -0.208 (0.151) -0.000 (0.000) 0.001 (0.001) 0.676** (0.317) -0.934** (0.407) 0.027 (0.020) 0.000 (0.000) -0.052 (0.119)
.1338 .0719 136 2.16 135
Lag 2 -0.278 (0.747) -0.171 (0.146) -0.000 (0.000) 0.003** (0.001) 0.878** (0.397) -0.681 (0.464) 0.022 (0.019) 0.000 (0.000) -0.265* (0.146) 0.693 (0.518) -0.225 (0.446) 0.088 (0.439) 0.119 (0.531) -0.059 (0.437) -0.130 (0.456) -0.227 (0.602) 0.378 (0.441) -0.080 (0.456) -0.233 (0.471) 0.677 (0.508) 0.223 (0.441) 0.196 (0.434) -0.085 (0.470) -0.874 (0.707) 0.119 (0.453) 0.756 (0.525) 0.251 (0.487)
Lag 1 0.476*** (0.105) -0.883** -0.681 (0.412) (0.451) 0.199*** 0.143* (0.066) (0.077) -0.000*** -0.000 (0.000) (0.000) -0.000 -0.000 (0.002) (0.002) 0.405*** 0.509*** (0.128) (0.142) -0.131 -0.343 (0.263) (0.294) -0.044*** -0.030* (0.014) (0.016) 0.000 0.000 (0.000) (0.000) -0.119 -0.308** (0.107) (0.132) -0.276 (0.501) -0.941** (0.444) -0.678 (0.438) -0.541 (0.512) -0.789 (0.440) -0.931** (0.459) -0.920 (0.583) -0.378 (0.446) -0.843* (0.462) -0.965** (0.461) -0.097 (0.555) -0.504 (0.446) -0.621 (0.436) -0.919** (0.457) -1.009* (0.592) -0.691 (0.445) -0.023 (0.496) 1.175 (0.450)
Mean 0.579*** (0.140) -1.271*** -1.009* (0.483) (0.526) 0.105 0.0686 (0.073) (0.069) -0.000 -0.000 (0.000) (0.000) 0.001 0.000 (0.002) (0.002) 0.878*** 1.019*** (0.210) (0.216) -1.040*** -1.170*** (0.380) (0.402) -0.014 -0.010 (0.009) (0.009) -0.000 -0.000 (0.000) (0.000) -0.143 -0.303** (0.128) (0.153) 0.366 (0.484) -0.111 (0.457) 0.050 (0.450) 0.175 (0.542) -0.085 (0.453) -0.238 (0.472) -0.138 (0.622) 0.291 (0.453) -0.174 (0.464) -0.199 (0.467) 0.649 (0.518) 0.242 (0.452) 0.154 (0.444) -0.220 (0.467) 1.019 (0.625) 1.080** (0.541) 0.050 (0.458) 0.686 (0.510) 0.528 (0.472)
.3399 .1825 136 2.16 109
.5370 .5044 138 16.49 128
.4504 .4163 155 13.20 145
.6463 .5635 138 7.80 111
.6042 .5201 155 7.18 127
77
Table 4.1.2 – Dependent Variable OIAT - Determinants of IT Outsoucing Intercept Return on Assets Return on Equity Real Revenue Revenue Growth Overhead Current Liabilities/Revenue Leverage Revenue/Employee Related Diversification Banking and Finance Professional Services Telecommunications Healthcare
0.021** (0.009) -0.048 (0.030) -0.007** (0.003) -0.000 (0.000) 0.000*** (0.000) 0.055*** (0.019) -0.059** (0.024) 0.002** (0.001) -0.000 (0.000) -0.004 (0.007)
Information Technology Media and Entertainment Insurance Electronics Energy Pharmaceuticals and Medical Equip Transportation Chemicals Manufacturing Retail and Distribution Utilities Hospitality and Travel Mining and Natural Resources Food and Beverage Consumer Goods R2 Adjusted R2 N F-value Df
.2319 .1723 126 3.89 116
Lag 3 -0.005 (0.023) -0.001 (0.002) -0.000 (0.000) 0.000 (0.000) 0.038** (0.016) -0.043** (0.019) 0.000 (0.000) -0.0000 (0.000) -0.007 (0.006) 0.039** (0.019) 0.002 (0.017) 0.007 (0.017) 0.005 (0.021) 0.001 (0.017) -0.008 (0.018) -0.005 (0.023) 0.015 (0.017) -0.001 (0.018) -0.007 (0.019) 0.226*** (0.027) 0.004 (0.017) 0.007 (0.017) -0.003 (0.018) -0.003 (0.023) 0.002 (0.018) 0.030 (0.020) 0.011 (0.019) .7240 .6515 126 9.99 99
0.029*** (0.009) -0.059 (0.042) -0.006 (0.009) -0.000 (0.000) 0.000 (0.000) 0.032 (0.018) -0.056** (0.023) 0.001 (0.001) -0.000 (0.000) -0.002 (0.007)
.1219 .0592 136 1.94 126
Lag 2 -0.039 (0.040) -0.006 (0.008) -0.000 (0.000) 0.000** (0.000) 0.038* (0.021) -0.019 (0.025) 0.001 (0.001) 0.000 (0.000) -0.015* (0.008) 0.065** (0.029) -0.010 (0.024) -0.004 (0.024) -0.003 (0.029) -0.004 (0.024) -0.005 (0.025) -0.012 (0.032) 0.014 (0.024) -0.005 (0.025) -0.016 (0.025) 0.067** (0.027) 0.008 (0.024) 0.005 (0.024) 0.004 (0.025) -0.048 (0.038) -0.003 (0.024) 0.024 (0.028) 0.013 (0.026)
Lag 1 0.028*** (0.005) -0.062*** -0.064*** (0.020) (0.021) 0.0256*** 0.020*** (0.003) (0.004) -0.000*** -0.000 (0.000) (0.000) -0.000 -0.000 (0.000) (0.000) 0.001 0.001 (0.006) (0.006) -0.003 -0.005 (0.012) (0.014) -0.006*** -0.005*** (0.001) (0.001) -0.000 -0.000 (0.000) (0.000) -0.004 -0.016** (0.005) (0.006) 0.064*** (0.023) -0.002 (0.020) 0.004 (0.020) 0.006 (0.024) 0.002 (0.020) -0.001 (0.021) -0.009 (0.027) 0.019 (0.021) -0.004 (0.021) -0.005 (0.021) 0.043* (0.026) 0.012 (0.021) 0.006 (0.020) 0.001 (0.021) -0.013 (0.027) 0.002 (0.021) 0.023 (0.023) 0.023 (0.021)
Mean 0.034*** (0.007) -0.067*** -0.056** (0.025) (0.027) 0.007* 0.004 (0.004) (0.004) -0.000 0.000 (0.000) (0.000) 0.000 0.000 (0.000) (0.000) 0.030*** 0.024** (0.011) (0.011) -0.071*** -0.058*** (0.019) (0.021) -0.001** -0.001 (0.001) (0.001) -0.000 -0.000 (0.000) (0.000) -0.002 -0.014* (0.007) (0.008) 0.051** (0.025) -0.005 (0.024) -0.004 (0.023) -0.008 (0.028) -0.006 (0.023) -0.011 (0.024) -0.015 (0.032) 0.007 (0.023) -0.014 (0.024) -0.008 (0.024) 0.061** (0.027) 0.005 (0.023) -0.000 (0.022) -0.010 (0.024) 0.010 (0.032) 0.008 (0.028) -0.010 (0.024) 0.012 (0.026) 0.034 (0.024)
.4015 .2588 136 2.81 109
.4452 .4062 138 11.41 128
.1353 .0816 155 2.52 145
.5972 .5028 138 6.33 111
.4031 .2762 155 3.18 127
78
Table 4.1.3 – Dependent Variable OIREV - Determinants of IT Outsoucing Lag 3 Constant Return on Assets Return on Equity Real Revenue Revenue Growth Overhead Current Liabilities/Revenue Leverage Revenue/Employee Related Diversification Banking and Finance Professional Services Telecommunications Healthcare
0.009 (0.006) -0.013 (0.019) 0.001 (0.002) -0.000 (0.000) 0.000** (0.000) 0.034*** (0.012) -0.019 (0.015) -0.000 (0.000) 0.000 (0.000) -0.004 (0.004)
Information Technology Media and Entertainment Insurance Electronics Energy Pharmaceuticals and Medical Equip Transportation Chemicals Manufacturing Retail and Distribution Utilities Hospitality and Travel Mining and Natural Resources Food and Beverage R2 Adjusted R2 N F-value Df
.1661 .1014 126 2.57 116
-0.012 (0.020) 0.002 (0.002) -0.000 (0.000) 0.000 (0.000) 0.038*** (0.013) -0.028* (0.016) -0.000 (0.000) -0.000 (0.000) -0.005 (0.005) 0.035** (0.016) 0.004 (0.014) 0.014 (0.014) 0.013 (0.017) 0.003 (0.014) -0.006 (0.015) -0.002 (0.019) 0.013 (0.014) 0.002 (0.015) -0.007 (0.016) 0.074*** (0.023) 0.006 (0.014) 0.007 (0.014) -0.004 (0.015) 0.001 (0.019) 0.005 (0.019) 0.012 (0.017) 0.006 (0.016) .4660 .3258 126 3.32 99
Lag 2 0.008 (0.006) 0.012 (0.027) -0.010* (0.006) -0.000 (0.000) 0.000 (0.000) 0.032*** (0.012) -0.023 (0.015) 0.001* (0.001) 0.000 (0.000) -0.002 (0.004)
.1365 .0749 136 2.21 126
-0.004 (0.026) -0.007 (0.005) -0.000 (0.000) 0.000** (0.000) 0.039*** (0.014) -0.025 (0.016) 0.001 (0.001) 0.000 (0.000) -0.010* (0.005) 0.054*** (0.018) -0.004 (0.015) 0.012 (0.015) 0.010 (0.018) -0.001 (0.015) -0.004 (0.016) -0.007 (0.021) 0.011 (0.015) -0.001 (0.016) -0.010 (0.016) 0.020 (0.017) 0.008 (0.015) 0.006 (0.015) -0.003 (0.016) -0.031 (0.024) 0.002 (0.016) 0.009 (0.018) 0.006 (0.017)
Lag 1 0.011*** (0.004) -0.011 -0.015 (0.016) (0.017) 0.006*** 0.003 (0.003) (0.003) -0.000*** -0.000 (0.000) (0.000) -0.000 -0.000 (0.000) (0.000) 0.023*** 0.025*** (0.005) (0.005) -0.001 -0.009 (0.010) (0.011) -0.001*** -0.001 (0.001) (0.001) 0.000 0.000 (0.000) (0.000) -0.002 -0.010** (0.004) (0.005) 0.031 (0.018) -0.019 (0.016) -0.008 (0.016) -0.005 (0.019) -0.018 (0.016) -0.023 (0.017) -0.023 (0.021) -0.007 (0.016) -0.019 (0.017) -0.028 (0.017) 0.005 (0.020) -0.011 (0.016) -0.015 (0.016) -0.023 (0.017) -0.025 (0.022) -0.018 (0.016) -0.013 (0.018) 0.0282 (0.016)
.4109 .2704 136 2.92 109
.7236 .7042 138 37.24 128
.8014 .7549 138 17.23 111
0.011** (0.005) -0.020 (0.017) 0.004 (0.003) -0.000 (0.000) 0.000 (0.000) 0.040*** (0.007) -0.025* (0.013) -0.001 (0.000) -0.000 (0.000) -0.004 (0.005)
.6949 .6760 155 36.70 145
Mean -0.030 (0.018) 0.003 (0.002) 0.000 (0.000) 0.000 (0.000) 0.045*** (0.008) -0.036** (0.014) -0.000 (0.000) -0.000 (0.000) -0.009* (0.005) 0.041** (0.017) 0.003 (0.016) 0.010 (0.016) 0.013 (0.019) -0.001 (0.016) -0.006 (0.016) -0.003 (0.022) 0.010 (0.016) -0.002 (0.016) -0.009 (0.016) 0.021 (0.018) 0.010 (0.016) 0.006 (0.016) -0.005 (0.016) 0.020 (0.019) 0.003 (0.016) 0.005 (0.018) 0.011 (0.016) .7814 .7349 155 16.81 127
79
4.2 Industry Effects Typically the literature does not address differences between industries with regard to IT outsourcing. Nonetheless, it is reasonable suspect that some industries may more inclined to outsource than others. The regressions suggest banking and financial services, transportation companies have higher outsourcing intensities. This may be due to so called bandwagon effects (also referred to previously as the Kodak Effect) where firms mimick the behavior of their peers. Alternatively, there may be certain industry characteristics such as comparatively centralized control systems which account for this. Airlines, for example, employ large centralized databases and reservations systems that, despite the increasing use of distributed information systems in other industries, continue to prove difficult to displace (Sheldon, 1997, O’Connor, 1995). At the other end of the spectrum professional service firms, information technology, energy, retail/distribution, hospitality/travel, metals/natural resources, and pharmaceuticals/medical equipment displayed lower outsourcing intensity than other types of industries. Factors such as tight profit margins, highly specialized R&D, or regulated environments may account for comparatively low levels of IT outsourcing activity. Other industries did not exhibit any pronounced patterns with regard to outsourcing intensity.
4.3 Outsourcing Effects While examining why firms choose to outsource is important to examine, equally
80
important is whether or not firm performance improves as a result of outsourcing. Data are much more limited in this regard, and a preliminary look at the impacts outsourcing using six measures referenced in the literature does not shed much light on the subject. These indicators deal with issues such as efficient use of assets, such as human and physical capital (revenue per employee, revenue divided by assets), earnings as a function of dollars invested (return on investment), earnings in comparison to revenues (return on sales), market valuations (market-to-book ratio) and year-over-year revenue growth. Perhaps because of the relatively small dataset available to examine (n = 124), there were few significant coefficients for outsourcing intensity, and the results were mixed. Outsourcing intensity appeared to be positively related to firm return on investment in the first year after the outsourcing event, but negatively related to marketto-book ratios one year following the event. No other outsourcing intensity coefficients were found to be significant with regard to the six measures of firm performance. Clearly however, this is an area ripe for further study. Suggested methodological approaches have detailed in Appendix C.
4.4 Other Findings In addition the hypotheses tested earlier with regard to degree of outsourcing intensity, the literature also suggests some other possible areas for exploration. The first is the belief that the term of outsourcing contracts has gotten shorter (Lacity 1998). Early deals, such as Kodak, were often for ten years, and it has been argued that such lengthy contracts increase the risk that the firm can be hostage to the outsourcer. As technology improves, outsourcers are able to take advantage of the efficiencies and economies, while
81
the customer or buyer is locked into a fixed-price contract, relegating most of the benefit to the outsourcer. In addition, business conditions often change, and long-term (5, 7, 10 year) outsourcing contracts are comparatively static vehicles, which can create a mismatch between outsourcer and customer as goals grow apart. Employment contracts, by comparison, are typically at-will arrangements, and severance packages (in the U.S.) cover much shorter time periods, which in turn give firms more flexibility. With regard to the analysis of functional areas outsourced, the dataset was reduced by two observations because the management consulting function was not tracked consistently by INPUT during the 1990-1999 timeframe. In fact, there is strong evidence to support the hypothesis that the term of outsourcing deals is becoming shorter as evidenced in Table 4.4.1. Even after controlling for other variables such as the type of function outsourced, or industry, the contract date coefficient is still highly significant an negative, thus lending support to the view that IT outsourcing contracts are becoming shorter. The alternative specifications include type of function outsourced, and industry. In each instance, the control variables were not significant and the main effect – the date in which the contract was signed – remains significant. In other words, the length of IT outsourcing contracts has decreased from 1990 – 1999.
82
Table 4.4.1 – IT Contract Term Contract Term Constant Contract Year
9.648 *** (0.540) -0.548 *** (0.073)
Application Management Application Operations Business Operations Desktop Services Electronic Markets Network/Carrier Management Platform Operations Systems and Technology Services Banking and Finance Professional Services Telecommunications Health Care Information Technology Media and Entertainment Consumer Goods Insurance Electronics Energy Pharmaceuticals and Medical Equipment Transportation Chemicals Manufacturing Retail and Distribution Utilities Hospitality Metals and Natural Resources Food and Beverage R2 Adjusted R2 N F-value Df
0.1604 0.1575 297 56.34 295
-0.351 *** (0.112) 8.349 (1.074) 1.179 (0.964) -0.560 (0.920) 0.218 (1.118) -0.162 (1.120) 0.470 (0.978) -0.471 (1.275) -0.465 (0.888) 0.1823 0.1596 297 8.03 288
-0.523 *** (0.076) 0.709 (2.779) -0.297 (2.831) -0.332 (2.785) -0.332 (3.015) -0.0295 (2.823) 0.229 (2.904) 9.184 (2.818) 0.364 (2.864) -1.050 (2.832) 0.491 (2.848) -0.087 (2.971) 1.514 (2.848) 0.430 (2.865) 0.7478 (2.792) 0.968 (2.857) 0.662 (2.919) -0.695 (3.014) 0.602 (2.849) 0.335 (3.079) 0.2013 0.1465 297 3.68 277
83
In addition, the changes in the term of IT outsourcing contracts over the course of the 1990s, it is worthwhile to examine whether the types of IT functions outsourced has changed also. Here again, there is evidence to support such a conclusion. Specifically, applications management, business operations, electronic markets, and systems and technology services have increased in prevalence over the past decade. By contrast, application operations and platform operations have decreased. The relative maturity of the technologies may be a causal factor. Electronic market outsourcing, for example, is not surprising given the substantial impact of the Internet during the 1990s, and the shortage of IT workers in such disciplines. Platform operations and applications operations, by contrast, are mature technologies and growth in both their use and opportunity for outsourcing has become relatively flat.
84
Table 4.4.2 – IT Contract Date Constant Application Management Application Operations Business Operations Desktop Services Electronic Markets Network/Carrier Management Platform Operations Systems and Technology Services Banking and Finance Professional Services Telecommunications Health Care Information Technology Media and Entertainment Consumer Goods Insurance Electronics Energy Pharmaceuticals and Medical Equipment Transportation Chemicals Manufacturing Retail and Distribution Utilities Hospitality Metals and Natural Resources Food and Beverage R2 Adjusted R2 N F-value Df
6.182 *** (0.434) -2.569 *** (0.485) 1.869 *** (0.473) -0.105 (0.590) 1.890 *** (0.580) 0.485 (0.515) -1.182 * (0.669) 2.000 *** (0.453) 0.5738 0.5635 297 55.58 289
Contract Date (Year of IT Contract) 5.942*** (1.513) -2.634*** (0.503) 1.854*** (0.483) 0.130 (0.616) 1.745*** (0.595) 0.557 (0.523) -1.076 (0.680) 2.058*** (0.680) -0.882 0.066 (2.183) (1.458) -0.118 0.671 (2.225) (1.488) -0.634 0.447 (2.188) (1.462) -1.400 0.925 (2.369) (1.589) -0.526 -0.020 (2.218) (1.481) -2.111 -0.143 (2.279) (1.524) -0.500 0.603 (2.250) (1.503) -0.882 -0.146 (2.225) (1.485) -0.929 -0.000 (2.238) (1.499) -0.167 0.702 (2.335) (1.560) -0.929 0.600 (2.238) (1.497) -1.25 0.322 (2.250) (1.507) -0.735 0.187 (2.194) (1.464) -1.385 -0.101 (2.244) (1.499) -1.125 -0.031 (2.293) (1.536) 0.200 0.241 (2.369) (1.579) -2.800 -0.423 (2.233) (1.496) -1.750 1.820 (2.417) (1.627) 0.075 0.5998 0.015 0.5629 297 297 1.25 16.24 278 271
CHAPTER FIVE SUMMARY AND POLICY IMPLICATIONS Research into the determinants of IT outsourcing by firms has focused largely on case study and cross-sectional approaches. Examination of IT outsourcing determinants, using publicly reported and audited data, as well as longitudinal data enables analysis of outsourcing in a more robust fashion, are explored here more fully than in many previous studies in an attempt to improve the external and construct validity of conclusions drawn. Outsourcing, and all of its many evolving variations up to and including joint ventures, is an organizational innovation that is comparable to the use of M-Form organizational structure adopted by large corporations early in the 20th century. In effect, outsourcing – in all of its many manifestations – may be again changing the nature of the dominant organizational form, in this case to the virtual organization.
5.1 Research Findings The dataset examined consists of 299 outsourcing events during the 1990-1999 timeframe, combined with financial data reported prior to the event. Several conclusions can be drawn from the analysis. First, evidence suggests that IT outsourcing by firms has increased during the past decade. Firms also appear to increase IT outsourcing intensity as a response to high overhead (sales, general and administrative) costs, or as a response to cash flow pressures. Further, banking and financial services, and transportation firms appear to
85
have higher outsourcing intensity than firms in other industry groups. Finally, there is strong evidence to indicate that firms are entering into IT outsourcing contracts for shorter periods of time apparently to avoid becoming hostage (i.e, the creation of approbriable quasi-rents) to the outsourcer and are able to maintain competitive pressures. There is also evidence of a shift in the types of IT functions outsourced over the decade, which may simply suggest that as the emphasis on and use of certain technologies shift, so does their tendency to be outsourced. Not surprisingly perhaps, there is a significant increase in the tendency to outsource electronic market (more commonly known as eCommerce) activities. Clearly this is due to the relative youth of such technologies, and the corresponding inability to staff such competencies within the firm as quickly as the market can. Because of the advantages that accrue to specialization in an information economy, firms focused on their core compentencies – both purchasers and providers of outsourcing services – are able to attract and reward such scarce talent as web developers. This contrasts with the early days of outsourcing where many firms built their own data centers and later outsourced them primarily to lower costs. As evidenced by the behavior of firms to outsource leading edge technologies, cost savings may now be more of a by-product or second-order effect of the decision to outsource than the main effect.
5.2 Public Policy Implications There are several links between the political economy literature and outsourcing that have important public policy implications, many of which must be addressed indirectly. As mentioned earlier, increased specialization is one of the means by which an
87
economy can become more productive. Yet, specialization drives asset specificity, which can cause holdup problems where the owner of an asset, either physical or intellectual, can charge monopoly rents. This in turn can cause economic resources to be allocated inefficiently and can reduce the ability of assets to be re-deployed to other uses, which has social welfare costs. In such scenarios, resources may be underutilized, involuntary unemployment can result, too many low-productivity units may be kept in production, and/or unbalanced job creation and destruction occurs (Caballero, 1998). Firms have typically reacted to asset specificity by vertically integrating, yet this introduces Xefficiency problems which stem in part from workers who are not exposed to market signals or lack incentives that would increase output (Liebenstein, 1966). Outsourcing can improve X-efficiency – and increasingly holdup problems that could result from asset specificity are mitigated in at least a couple of ways by firms. One is the increasing insistence on the use of standardized, hardware independent IT platforms and protocols such as Java and TCP/IP (Internet standards). Another is through contractual innovations which are detailed below. Research presented in Chapter 4 suggests, for example, a trend in the 1990s toward the use of shorter term outsourcing agreements. Public policy implications of allocative efficiency, in addition to those of the social welfare cost of contractual bilateral monopoly embodied in outsourcing agreements, are also germane. Specifically, there is also the issue of trade policy and globalization, which has an impact on outsourcing. Restrictions on trade, for example, keep economic resources from being most efficiently utilized. Both restrictions on trade, as well as many social welfare “inefficiencies”, however, are in place not as the result of market competition, but because of other countervailing influences, such as power
88
relations and institutional factors. These countervailing influences have constituencies just as the market does, and serve to provide a degree of stability, whether for good or ill (Baker, et al. 1998). Labor unions, for example, are one such force which attempts to correct for perceived imbalances in the natural order of the market. Still, outsourcing of both tangible goods and services, whether in the U.S. or across its borders, continues at a brisk pace. Interestingly, while outsourcing is often blamed for depressing wages, as firms move operations outside of the U.S., the impact of outsourcing on wage levels is less than half than that of technology (Feenstra 1999). Competition, enabled by technology, pokes and prods until it finds an opening, as is evidenced for example in the telecommunications sector where wireless technology is beginning to bypass traditional copper wire based local loops or exchanges. The shift toward outsourcing in recent years, most notably in the service sector, has been largely driven by manufacturing firms which have been shedding nonproduction workers of all types (Feenstra, 1999; Fixler and Siegel, 1999). In addition to the ubiquitous use of outsourcing for non-production functions by manufacturing firms, service firms are beginning to outsource to a greater degree as well. Outsourcing of IT services, for example, is widespread in financial, insurance, real estate and transportation sectors (Fixler and Siegel, 1999). Firms are attracted to outsourcing in part because outsourcing provides the means to more closely match wage levels with the labor supply/demand function. Within firms or within industries, workers can be insulated from market signals. Outsourcing tends to expose middle managers whose job functions are redundant, or employees who wage rates are above market.
89
Perhaps surprisingly, the result of outsourcing on wage levels, partially across international borders, is mixed. While political rhetoric bemoans wage losses stemming from the use of outsourcing, high-skilled workers often benefit. Knowledge workers are highly specialized and have often improved their prospects as manufacturing industries have shed non-production jobs (Sharpe, 1997). This is because outsourcing has tended to increase the demand for skilled labor. Further, the impact of outsourcing on wages is predominantly within industries, not across them (Feenstra, 1996). In other words, it is not certain industries per se (e.g. agriculture) that receive the brunt of the impact from outsourcing, but rather particular job functions across all industries. The evidence also suggests that outsourcing makes the demand for skilled labor less elastic, while demand for unskilled labor becomes more elastic because of the prospect for moving operations offshore (Davis, 1999). With skilled labor, this is more difficult because of coordination issues, specialization, decreased opportunities for substitution and a limited supply. Though outsourcing is presumed to be the culprit, societal cleavages may ultimately have less to do with outsourcing, which is simply a symptom economic restructuring, than with the growing chasm between low and high skilled (typically non-production) workers. Outsourcing is merely one of the market mechanisms by which the relative demand for different skill sets and capabilities are matched with the available supply. More fundamentally, outsourcing can provide workers an alternative to traditional corporate hierarchies. Functional expertise, rather than career options based on industry focus, is the basis for the concept of outsourcing. Such functional specialties, which almost invariably cross over industry boundaries, now command equivalent importance in the marketplace. Liberation from hierarchy, or at least the creation of a greater number of
90
hierarchies from which choose, paves the way for greater individual freedom and choice:
…[M]ost of the modern tenets of liberal democracy actually stem from the classical liberalism of a group of writers and thinkers who, from the seventeenth century onward, argued that society should be based on the rights of individuals rather than on the hierarchy of a “great chain of being,” stretching from God downward, with everyone else assigned a fixed place. (Micklethwait and Wooldridge, 2000: 337)
Outsourcing chips away at hierarchy because it, by definition, relies on the market, not non-market institutions. When regulated industries or governments choose to outsource, they are in effect, attempting to impose market discipline on non-market institutions. As a governance mechanism, outsourcing offers firms an opportunity to increase the use of specialization as a way to increase economic efficiency and improve performance. Firms engage in outsourcing to acquire critical capabilities more quickly than would be possible through in-house development. Firms also outsource to restructure themselves from an organizing logic based on the industrial era to an organizing logic based on the information age (Lee and Venkatraman, 1999). On the flipside, outsourcers are able to take advantage of economies that individual firms cannot because of their ability to aggregate demand pooled across many firms for the performance of specific functions. And finally, individual freedom can be enhanced, because workers have a greater number of potential options than in a comparatively hierarchical institutional structures typically emblematic of traditional firms or governments. At the extreme, one might infer that outsourcing produces the end of the firm as we know it. Coase’s original argument suggests as much. The key thing to remember,
91
however, is that the firm remains a viable governance mechanism, even if its form undergoes significant transformation. Coase (1937) himself noted that the price mechanism is not always preferable; further, transactions organized within the firm are treated differently by governments and regulatory bodies. Firms appear likely to be with us for some time – they are merely being refined. Hence, the research results in this study strongly suggest that outsourcing in and of itself is not a panacea. The private sector continues to experiment with organizational innovation and uses the market (i.e. outsourcing) to operate more efficiently and effectively. As mentioned earlier, one of the ways in which this was manifested in the research results is in the tendency toward shorter-term contracts. Such innovative contracting techniques – which rely on improved ex ante definition of services to be delivered, as well as even-handed termination clauses – have application to public policy. Take, for example, the awarding of franchising rights for community cable television systems. Previous research has focused on ex ante competitive bidding as a means to ensure favorable terms in situations that were perceived to require long-term contracts (Zupan, 1989). Yet experience from the private sector, as well as Williamson’s (1985) identification of the “fundamental transformation” to ex post bilateral monopoly after contract signing, suggests that use of shorter term contracts and options for re-bid could be more effective tools to manage cable franchisees. In the private sector, even with high initial investments in infrastructure by outsourcers, large firms are able to incorporate contractual language that addresses termination for convenience, termination for cause, and renewal. These out-clauses can specify, for example, that in the case of termination or re-bid, the outsourcer will be reimbursed for the un-amortized portion of
92
their investment. This approach, if applied to community cable television and other public infrastructure bidding situations, could improve the responsiveness and performance of incumbent providers, while at the same time affording them protection for their substantial investment. However it should be noted that there is a clear onus on the public entity seeking to contract for services in that it be unambiguous about the definition of service, delivery and measurement, nor naïve about the costs. This is precisely why long-term contracts are inherently problematic: they are either necessarily incomplete or ultimately inflexible in the face of a changing political-economic environment. Still, while some level of sophistication is required, the issues are not insurmountable. In the final analysis, the lack of such sophistication by the contracting party, either public or private, should not result in a wholesale condemnation of outsourcing as a governance mechanism.
5.3 Suggestions for Future Research In the best of all possible worlds, standardized SEC reporting information on outsourcing activities of firms would be reported along with other financial data . Since such data are not currently available, the second best solution would be for research firms such as INPUT to track firm outsourcing activity and perhaps expand coverage to include non-IT activities (i.e . the remaining 60% of service outsourcing). Across the still limited range of academically oriented outsourcing literature, the universal grievance is with the paucity of available data with which to work. In addition, while there is room to further validate the reasons why firms choose to outsource in the first place, the real question will increasingly be: what results have
93
firms experienced by using outsourcing as a governance mechanism? More research is clearly needed in this area. Details of a proposed methodological approach to do so are included in Appendix C. In many cases, outsourcing has not fulfilled expectations. And if in fact firm (or government) performance cannot systematically be improved through the use of outsourcing, then the questions may not revolve around whether outsourcing helps improve firm performance, but rather what are the factors that do influence it, such as good management – in which case the decision to outsource would simply be an incidental event.
APPENDICES
94
95
Appendix A – IT Outsourcing Determinants Sample Statistics
Industry Category Banking and Finance Professional Services Telecommunications Healthcare Information Technology Media and Entertainment Consumer Goods Insurance Construction and Engineering Electronics Energy Pharmaceuticals and Medical Equipment Transportation Chemicals Manufacturing Retail and Distribution Utilities Hospitality and Travel Metals and Natural Resources Food and Beverage Processing Totals
Function Outsourced Systems and Technology Services Business Process Management Application Operations Network/Carrier Management Electronic Markets Desktop Services Application Management Business Operations Platform Operations Management Consulting Totals
No. of Firms 52 17 41 5 19 9 1 12 0 18 14 5 14 12 35 13 8 5 15 4 299
# of Events 121 50 44 27 14 13 11 9 8 2 299
% of Firms 17.4 5.7 13.7 1.7 6.4 3.0 0.3 4.0 0.0 6.0 4.7 1.7 4.7 4.0 11.7 4.7 2.7 1.7 5.0 1.3 100%
% of Events 40.5 16.7 14.7 9.0 4.7 4.3 3.7 3.0 2.7 0.7 100%
96
Appendix B – Outsourcing Effects Methodological Background
Suggestions for Future Research – Outsourcing Effects Having explored the relationships between financial indicators and the tendency to outsource, the next logical step would be to examine what in fact the results from IT outsourcing have been. Such an analysis is the subsequent leg of an extended causal sequence linking outsourcing intensity to economic or financial outcomes. In this second phase of analysis, outsourcing intensity becomes the independent variable, and measures of firm performance are now introduced as dependent variables. It is in this fashion that it is hoped that the consequences of outsourcing intensity can be examined. Given the increasing frequency of outsourcing as a corporate governance mechanism, and having attempted to answer questions about what measures predict the likelihood to outsource, it is only natural to wonder how outsourcing has affected firm performance. To what extent has the promise been realized? Given that outsourcing is a relatively young tool for corporate management, there are only early returns available for examination, so to speak. Nonetheless enough data should be available to draw some preliminary conclusions. Dependent Variable The literature has generally identified six indicators of performance by firms (Mahmood and Mann, 1993):
Revenue/Employees: sales divided by the number of employees (labor productivity) Revenue/Assets: sales divided by assets
Return on Investment: earnings before income taxes, depreciation
97
anamortization divided by invested capital Return on Sales: earnings before income taxes, depreciation and amortization divided by sales Market-to-Book Ratio: market value divided by book value Revenue Growth: year-over-year increase in sales
These indicators deal with issues such as efficient use of assets, such as human and physical capital (revenue per employee, revenue divided by assets), earnings as a function of dollars invested (return on investment), earnings in comparison to revenues (return on sales), market valuations (market-to-book ratio) and year-over-year revenue growth. Research into the results from outsourcing have been almost exclusively based on survey data. To a large extent, this has been because the results from outsourcing have been difficult to quantify and many exogenous causal variables are believed to play a role in a firm’s success or failure. Nonetheless, this should not deter attempts to uncover systematic effects of the impact of outsourcing on firm performance. This dissertation represents a first step toward such a measure using publicly available data on firm activity.
Predictors and Covariates Firm performance can be impacted by factors other than IT outsourcing, such as ownership structure (Jensen and Meckling, 1976), degree of automation (Harris and Katz, 1989), quality of management, stage of the business cycle, revenue growth and labor productivity (Frydman, et al., 1999) and the level of competition in the market or a given industry (Mahmood and Mann, 1993). In order to isolate other possible causes for firm performance, proxies for the quality of firm management are used to attempt to assess whether changes in firm performance are caused by outsourcing, or whether management
98
quality is the proximate cause, and outsourcing merely a by-product.
Degree of Outsourcing H7: Outsourcing intensity and firm performance are positively related. In the previous section, outsourcing intensity was introduced as the dependent variable. Given that firms use outsourcing as a mechanism to improve performance, it will be worthwhile to explore whether there is any evidence that in fact such is the case. As before, outsourcing can be normalized in a variety of ways. Ideally, we would have access to the IT budgets for firms, but for a variety of reasons, such data are unavailable. Therefore I have used to revenues and assets to normalize data on outsourcing intensity in addition to an estimate of the IT budget based on the industry in which the firm operates.
IT outsourcing divided by revenues IT outsourcing divided by assets IT outsourcing divided by estimated IT budget
Cost Control (Quality of Management) H8a: Quality of management (use of cost controls) and firm performance are positively related. Firm firm performance can be the result of factors other than the degree to which it outsources one of more functions. Further, outsourcing may simply be an artifact of good management technique, which draws from many disciplines, but manifests itself in other ways. It will be important to test for these possible relationships as well. One such measure is the ability of management to control costs effectively. Measures of overhead cost structures include:
99
Overhead(1): Sales, general and administrative expenses (SG&A) divided by revenues Overhead(2): Cost of goods sold plus SG&A divided by revenues
Cash Flow (Quality of Management)
H8b: Firm management quality (ability to generate cash flow) and firm performance are positively related.
Other measures of management quality include the ability to maintain sufficient capital and liquidity to conduct business and fund new investment. Five measures have been identified:
Cash Flow(1): Cash Flow(2): Cash Flow(3): Cash Flow(4): Cash Flow(5):
Cash and cash equivalents divided by revenue. Total liabilities divided by revenues Long-term debt divided by revenues Current liabilities divided by revenues Total liabilities divided by shareholder equity
The hypothesis to be tested is similar to H8a but with different measures. Experiential Effects and Dummy Variables
H9: Recent outsourcing and firm performance are positively related. In order to control differences in outsourcing and firm performance characteristics across industries, dummy variables have been included in the model in order to isolate these effects. In addition, because the reasons for outsourcing may have changed over time, “experience” variables have been included which match the outsourcing event to the period of time in which it occurred. The INPUT database of outsourcing events covers the period 1990-99, and in both analyses, dummy independent variables have been
100
included in order to examine these effects. It may be that firms who outsourced more recently have learned from the mistakes of others that have gone before.
Summary of Variables The variables to be examined for outsourcing performance can be summed up in the following tables.
101
Outsourcing Effects Variables Dependent Variable: Code Indicator PI Performance of Firm after Decision to Outsource
Independent Variables: Outsourcing Variable Name Determinants 1. Outsourcing Intensity
2. Cost Focus
3. Cash Needs
Operational Definition Six indicators of firm performance: -Revenue/Employees -Revenue/Assets -Return on Investment -Return on Sales -Market-to-Book Ratio -Revenue Growth
Predicted Relationship
Data Source CompuStat Database
Variable Code
Operational Definition
Data Source
Outsourcing normalized by Revenues Outsourcing normalized by assets
+
OIREV
Annual amount of IT outsourcing deal divided by revenues
Compustat
+
OIASST
Annual amount of IT outsourcing deal divided by assets
Compustat
Outsourcing normalized by estimate of IT budget Overhead expenses
+
OIIT
Annual amount of IT outsourcing deal divided by estimated IT budget
Compustat
-
OH1
Sales, general and administrative expenses (SG&A) divided by revenues
Compustat
Total expenses
-
OH2
Cost of goods sold plus SG&A divided by revenues
Compustat
Liquid assets on hand
-
CF1
Cash and cash equivalents divided by revenues
Compustat
Ability to service debt
-
CF2
Total liabilities divided by revenues
Compustat
Ability to service debt
-
CF3
Long-term debt divided by revenues
Compustat
Ability to service debt
-
CF4
Current liabilities divided by revenues
Compustat
Financial leverage
+
CF5
Total liabilities divided by shareholder equity
Compustat
The applicable causal diagram and explanations for the operationalized variables for outsourcing determinants are outlined in Figure 3.1. Y = f (X 1 , X 2 , X 3 , …, X 13 )
102
Causal Diagram
Outsourcing
Cost Control
Cash Flow
Revenue per Employee Revenue divided by Assets Return on Sales Return on Investment Market-to-Book Ratio Revenue Growth
Control Variables
Outsourcing Intensity X 1 = Annual outsourcing divided by IT budget X 2 = Annual outsourcing divided by revenue X 3 = Annual outsourcing divided by assets
Cash needs X 6 = Liquidity (cash and cash equivalents divided by revenues) X 7 = Ability to service debt (total liabilities divided by revenues) X 8 = Ability to service debt (long-term debt divided by revenues) X 9 = Ability to service debt (current liabilities divided by revenues) X 10 = Financial leverage
Cost focus X 4 = Overhead expenses X 5 = Total expenses
Control Variables X 11 = Annual year-over-year GDP growth X 12 = Experience dummy variable (year in which outsourcing contract was announced) X 13 = Firm size as measured by real revenues
Outsourcing Effects The suggested regression equations for measuring outsourcing effects are as follows: PI = β 0 + β 1 OH1 + β 2 OH2 + β 3 CF1 + β 4 CF2 + β 5 CF3+ β 6 CF4 + β 7 CF5 + β 8 ECON+ β 9 EXP + β 10 RREV + β 11 OIIT+ β 12 OIREV + β 13 OIAT + µ i
Where PI = Performance Indicator. Six performance indicators have been
103
identified in the literature: Revenue/Employees Revenue/Assets Return on Investment Return on Sales Market-to-Book Ratio Revenue Growth OI = Outsourcing intensity OH1 = SG&A divided by revenues OH2 = CGS plus SG&A divided by total revenue CF1 = Cash and equivalents divided by revenues CF2 = Total liabilities divided by revenues CF3 = Long-term debt divided by revenues CF4 = Current liabilities divided by revenues CF5 = Total liabilities divided by common shareholder equity ECON = Year over year growth in U.S. GDP EXP = Experience dummy variable (Year in which outsourcing contract was announced) RREV = Inflation-adjusted revenues OIIT = Annual IT outsourcing expenditure divided by estimated total IT budget OIREV = Annual IT outsourcing expenditure divided by revenue OIAT = Annual IT outsourcing expenditure divided by total assets
104
REFERENCES Alchian, Armen A. and Harold Demsetz. “Production, Information Costs, and Economic Organization.” American Economic Review, 62 (December 1972), 777-95. Ang, Soon and Detmar W. Straub. “Production and Transaction Economies and IS Outsourcing: A Study of the U.S. Banking Industry,” MIS Quarterly 22: 4 (December 1998) 535-52. Arnett, Kirk P. and Mary C. Jones. “Firms That Choose Outsourcing: A Profile,” Information and Management 26 (1994) 179-88. Austrian, Geoffrey D. Herman Hollerith: Forgotten Giant of Information Processing. New York: Columbia University Press 1982. Baker, Wayne E. and Robert R. Faulkner, Gene A. Fisher. “ Hazards of the Market: The Continuity of Interorganizational Market Relationships,” American Sociological Review, 63 (April 1998) 147-77. Brown, Robert M., Amy W. Gatian and James O. Hicks, Jrs. “Strategic Information Systems and Financial Performance”, Journal of Management Information Systems, 11: 4 (Spring 1995) 215-48. Brynjolfsson, Erik. “Information Assets, Technology and Organization,” Management Science. 40: 12 (December 1994), pp. 1645-62. Caballero, Ricardo J. “The Macroeconomics of Specificity,” Journal of Political Economy, 106:4 (1998) 724-67. Chandler, Alfred D. Strategy and Structure: Chapters in the History of American Industrial Enterprise. Cambridge, MA, M.I.T. Press (1962). Chaudhury, A., Kichan Nam and H. Raghav Rao. “Management of Information Systems Outsourcing: A Bidding Perspective,” Journal of Management Information Systems, 12: 2 (Fall 1995) 131-59. Cheung, Steven N.S. “The Transaction Costs Paradigm: 1998 Presidential Address, Western Economic Association,” Economic Inquiry, 36 (October 1998) 514-21. Clark, Thomas and Robert Zmud, Gordon McCray. “The Outsoucing of Information Services: Transforming the Nature of Business in the Information Industry,” Strategic Sourcing of Information Systems, edited by L.P. Willcocks and M.C. Lacity. New York: John Wiley & Sons, Ltd (1998).
105
Cross, John, Michael J. Earl and Jeffrey L. Sampler. “Transformation of the IT Function at British Petroleum,” MIS Quarterly 21:4 (December 1997) 401-23. Coase, Ronald. “The Nature of the Firm,” Economica, 4 (November 1937).386-405. Davis, John B. “Is Trade Liberalization an Important Cause of Increasing U.S. Wage Inequality? The Interaction of Theory and Policy,” Review of Social Economy, 57:4 (December 1999) 488-506. Deavers, Kenneth L. “Outsourcing: A Corporate Competitiveness Strategy, Not a Search for Low Wages,” Journal of Labor Research, 18:4 (Fall 1997) 503-19. Dewan, Sanjeev; Steven C. Michael and Chunk-ki Min. “Firm Characteristics and Investments in Information Technology: Scale and Scope Effects,” Information Systems Research, 9:3 (September 1998) 219-32. DiRomualdo, Anthony and Vijay Gurbaxani. “Strategic Intent for IT Outsourcing,” Sloan Management Review, 39: 4 (Summer 1998) 67-80. Dyer, Jeffrey H. “Effective Interfirm Collaboration: How Firms Minimize Transaction Costs and Maximize Transaction Value,” Strategic Management Journal, 18:7 (1997) 535-56. Downes, Larry and Chunka Mui. Unleashing the Killer App: Digital Strategies for Market Dominance, Boston: Harvard Business School Press (1998). Earl, Michael J. “The Risks of Outsourcing IT”, Sloan Management Review, 37 (Spring 1996) 26-32. England, Robert Stowe. “Take Part of Me: How Companies are Unlocking Value by Carving Out Pieces of their Business,” CFO Magazine, 15: 3 (March 1999) 97-99. Fama, Eugene F. “Agency Problems and the Theory of the Firm,” Journal of Political Economy, 88:2 (1980) 288-307. Feenstra, Robert C. and Gordon H. Hanson. “Globalization, Outsourcing and Wage Inequality,” American Economic Review, 86 (May 1996) 240-5. Feenstra, Robert C. and Gordon H. Hanson. “ The Impact of Outsourcing and HighTechnology Capital on Wages: Estimates for the United States, 1979-1990,” Quarterly Journal of Economics, 114:3 (August 1999) 907-40. Fixler, Dennis J. and Donald Siegel. “Outsourcing and Productivity Growth in Services,” Structural Change and Economic Dynamics, 10:2 (June 1999) 177-94.
106
Frydman, Roman, and Cheryl Gray, Marek Hessel, Andrej Rapaczynski. “When Does Privatization Work? The Impact of Ownership on Corporate Performance in the Transition Economies,” Quarterly Journal of Economics, 114:4 (November 1999) 115391. Gordon, John Steel. “Give Credit Where It’s Due,” Wall Street Journal. (May 25, 2000) A26. Grover, Varun, Myun Joong Cheon and James T.C. Cheng. “The Effect of Service Quality and Partnership on the Outsourcing of Information Systems Functions”, Journal of Management Information Systems, 12:4 (Spring 1996) 89-116. Guelpa, Fabrizio. “Corporate Governance and Contractual Governance: A Model” . Rivista Internazionale di Scienze Sociali, 2 (1998) 73-90. Guajarati, Damdar N. Basic Econometrics 3rd Ed. New York: McGraw-Hill (1995: 497539). Harris, S. E. and J. L. Katz. “Predicting Organizational Performance Using Information Technology Managerial Control Ratios,” Proceedings of the Twenty-Second Annual Hawaii International Conference on Systems Sciences, 4 (January 1989) 197-204. Hill, Charles W. L. “Internal Capital Market Controls and Financial Performance in Multidivisional Firms,” Journal of Industrial Economics, 37:1 (September 1988) 67-83. Jacquemin, Alexis P. and Charles H. Berry. “Entropy Measure of Diversification and Corporate Growth,” Journal of Industrial Economics, 27:4 (June 1979) 359-369. Jensen, Michael C. and William H. Meckling. “Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure,” Journal of Financial Economics, 3 (1976) 30560. Kettinger, William J., Varun Grover, Subashish Guha, and Albert Segars. “Strategic Information Systems Revisited: A Study in Sustainability and Performance,” MIS Quarterly, 18:1 (March 1994) 31-58. Kivijarvi, Hannu and Timo Saarinen. “Investment in Information Systems and the Financial Performance of the Firm,” Information and Management, 28 (1995) 143-63. Klein, Benjamin, Robert G. Crawford and Armen A. Alchian. “Vertical Integration, Appropriable Rents, and the Competitive Contracting Process,” Journal of Law and Economics. 21 (October 1978) 297-326. Lacity, Mary. An Interpretive Investigation of the Information Systems Outsourcing Phenomenon. University of Houston, (1992) Doctoral Dissertation.
107
Lacity, Mary C., Leslie P. Willcocks and David F. Feeny. “IT Outsourcing: Maximize Flexibility and Control,” Harvard Business Review, 73, (May-June 1995) 84-93. Lacity, Mary C., Leslie P. Willcocks and David F. Feeny. “The Value of Selective IT Sourcing”, Sloan Management Review, 37:3 (Spring 1996) 13-25 Lacity, Mary and Leslie Willcocks. “An Empirical Investigation of Information Technology Sourcing Practices: Lesson From Experience,” MIS Quarterly, 22: 3 (September 1998) 363-408. Lee, Chi-Hyon and N. Venkatraman. “New Organizational Arrangements for Information Technology Resource Leverage: Empirical Test of Value Capture,” Working Paper, Boston University (1999). Lee, J.A.N. Computer Pioneers. Los Alamitos, CA: IEEE Computer Society Press 1995. Leibenstein, Harvey. “Allocative Efficiency vs. ‘X-Efficiency’,” American Economic Review. 56:3 (June 1966) 392-414. Levin, Doron P. Irreconcilable Differences: Ross Perot versus General Motors. New York: Penguin Books (1989). Levy, David T. “The Transactions Cost Approach to Vertical Integration: An Empirical Examination,” Review of Economics and Statistics, (1985) 438-45. Loh, Lawrence and N. Venkatraman, “Determinants of Information Technology Outsourcing: A Cross-Sectional Analysis,” Journal of Management Information Systems, 9:1 (Summer 1992) 7-24. Loh, Lawrence and N. Venkatraman. “ Diffusion of Information Technology Outsourcing: Influence Sources and the Kodak Effect,” Information Systems Research, 3: 4 (December 1992) 334-58. Maddigan, Ruth J. “The Measurement of Vertical Integration,” Review of Economics and Statistics, 63:3 (1981) 328-35. Mahmood, Mo Adam and Gary J. Mann. “Measuring the Organizational Impact of Information Technology Investments: An Exploratory Study,” Journal of Management Information Systems, 10:1 (1993) 97-122. Malhotra, Yogesh. “An Empirical Analysis of the Determinants of Information Systems Productivity and the Role of Outsourcing Policy,” Brint Research Institute (1995)
108
McFarlan, F. Warren and Richard L. Nolan. “How to Manage an IT Outsourcing Alliance,” Sloan Management Review, 36:2 (Winter 1995) 9-23. Micklethwait, John and Adrian Wooldridge. A Future Perfect: The Challenge and Hidden Promise of Globalization. New York: Crown Publishers (2000). Mitra, Sabyasachi and Antoine Karim Chaya. “Analyzing Cost-Effectiveness of Organizations: The Impact of Information Technology Spending,” Journal of Management Information Systems, 13:2 (Fall 1996) 29-57. Norton, Seth W. “Information and Competitive Advantage: The Rise of General Motors,” The Journal of Law and Economics, 40 (April 1997).245-60. O’Connor, William E. An Introduction to Airline Economics, Westport, CT: Praeger Publishers (1995). Ohlson, James.A. “Financial Ratios and the Probabilistic Prediction of Bankruptcy,” Journal of Accounting Research, 18:1 (Spring 1980) 109-31. Ou, Jane A. and Stephen H. Penman “Financial Statement Analysis and the Prediction of Stock Returns,” Journal of Accounting and Economics, 11 (November 1989) 295-329. Oxley, Joanne E. “Appropriability Hazards and Governance in Strategic Alliances: A Transaction Cost Approach,” Journal of Law, Economics, and Organization, 13:2 (October 1997) 387-409. Palepu, Krishna G. “”Predicting Takeover Targets: A Methodological and Empirical Analysis,” Journal of Accounting and Economics, 8 (1986) 3-35. Palepu, Krishna. “Diversification Strategy, Profit Performance and the Entropy Measure,” Strategic Management Journal, 6 (1985) 239-55. Palvia, Prashant C. “A Dialectic View of Information Systems Outsourcing: Pros and Cons,” Information and Management, 29 (1995) 265-75. Poppo, Laura and Todd Zenger. “Testing Alternative Theories of the Firm: Transaction Cost, Knowledge-Based, and Measurement Explanations for Make-or-Buy Decisions in Information Services,” Strategic Management Journal, 19 (1998) 853-77. Prahalad, C.K. and Gary Hamel. “The Core Competence of the Corporation,” Harvard Business Review, 90:3 (1990) 79-91. Quinn, James Brian and Frederick G. Hilmer. “Strategic Outsourcing”, Sloan Management Review 35:4 (Summer 1994) 19-31.
109
Ricardo, David. The Principles of Political Economy and Taxation. London: Guernsey Press (1973) Saarinen, Timo and Ari P.J. Vepsalainen. “Procurement Strategies for Information Systems”, Journal of Management Information Systems, 11:2 (Fall 1994) 187-208. Segars, Albert H. and Varun Grover. “The Industry-Level Impact of Information Technology: An Empirical Analysis of Three Industries,” Decision Sciences, 26:3 (1995) 337-68. Sharpe, Murem. “Outsourcing, Organizational Competitiveness, and Work,” Journal of Labor Research, 18:4 (Fall 1997) 535-49. Sheldon, Pauline J. Tourism Information Technology, New York: CAB International (1997). Simon, Herbert. Administrative Behavior, 3rd Ed., The Free Press: Macmillan Publishing, New York (1976) Smith, Adam. An Inquiry into the Nature and Causes of the Wealth of Nations. Ann Arbor: Oxford University Press (1979) Smith, Michael Alan, Sabyasachi Mitra and Sridhar Narasimhan. “Information Systems Outsourcing: A Study of Pre-Event Firm Characteristics,” Journal of Management Information Systems, 15:2 (Fall 1998) 61-93. Spiller, Pablo T. “On Vertical Mergers,” Journal of Law, Economics and Organization, 1:2 (Fall 1985) 285-312. Teece, David. “Internal Organization and Economic Performance: An Empirical Analysis of the Profitability of Principal Firms,” Journal of Industrial Economics, 30:2 (December 1981) 173-99. Teng, James T.C., Myun Joong Cheon and Varun Grover. “Decision to Outsource Information Systems Functions: Testing a Strategy-Theoretic Discrepancy Model,” Decision Sciences, 26:1 (1995) 75-103. van der Meer-Kooistra, Jeltje and Ed G.J. Vosselman. “Management Control of Interfirm Transactional Relationships: The Case of Industrial Renovation and Maintenance,” Accounting, Organizations and Society, 25 (January 2000) 51-77. Venkatraman, N. “Beyond Outsourcing: Managing IT Resources as a Value Center,” Sloan Management Review, 38:3 (Spring 1997) 51-64. Venkatraman, N. and John C. Henderson. “Real Strategies for Virtual Outsourcing”,
110
Sloan Management Review, 40:1 (Fall 1998) 33-48. West, Lawrence A., Jr. “Researching the Costs of Information Systems”, Journal of Management Information Systems, 11:2 (Fall 1994) 75-107. Williamson, Oliver. Markets and Hierarchies. Analysis and Antitrust Implications. New York: Free Press, Simon and Schuster (1975). Williamson, Oliver. The Economic Institutions of Capitalism: Firms, Markets and Relational Contracting. New York: Free Press (1985). Williamson, Oliver. Economic Organization. Washington Square, New York: New York University Press (1986). Williamson, Oliver. Economics and Organization: A Primer” California Management Review, 38:2 (Winter 1996a) 131-61. Williamson, Oliver. The Mechanisms of Governance. New York: Oxford University Press (1996b). Zupan, Mark A. “The Efficacy of Franchise Bidding Schemes in the Case of Cable Television: Some Systematic Evidence” Journal of Law and Economics, 22 (October 1989) 401-456. *
*
*
VITA Thomas Nelson Tunstall was born in New Iberia, Louisiana, on October 6, 1956, the son of Karl Nelson Tunstall and Betty Jane Mestayer. After completing college preparatory work at R.L. Turner High School, Carrollton, Texas, in 1974, he entered East Texas State University at Commerce, Texas. In 1975 he transferred to the University of Texas at Austin. He received the degree of Bachelor of Business Administration with a major in Marketing from the University of Texas in August 1979. For approximately ten years, he was employed in operations and accounting roles in the hospitality industry with such companies as Marriott and Wyndham Hotels. In August of 1985, he married Renee Lynn Malkow of Oconomowoc, Wisconsin. They now have four children, Matthew Nelson (born 1987), Rachel Lauren (born 1988), Taylor Nicole (born 1995), and John Thomas (born 1996). In 1990 he entered the Graduate School of Management at The University of Texas at Dallas and completed an M.B.A. in 1992. In 1995 he entered the Graduate School of Social Sciences (Political Economy) of The University of Texas at Dallas to begin work on his doctorate. He was employed by American Airlines – SABRE for nearly ten years from 1989 – 1998 and is currently employed with KPMG, LLC (formerly Peat Marwick) as a senior consultant.
