Determinants of Information Technology Outsourcing: A Cross-Sectional Analysis

LAWRENCE LOH AND N. VENKATRAMAN

is a doctoral candidate in the Sloan School of Management at Massachusetts Institute of Technology. He is also a senior tutor in the Faculty of Business Administration at National University of Singapore. His current research interests include information technology outsourcing, corporate governance, and diffusion of innovation. His research has appeared in the Journal of Business Research and he has presented papers at the Academy of Management National Meeting and the TIMS/ORSA Joint National Meeting.
LAWRENCE LOH

N. VENKATRAMAN is Richard S. Leghorn Career Deveiopment Associate Professor of Management in the Sloan School of Management at Massachusetts Institute of Technology. His current research interests intersect strategic management and information technology. At Sloan, he teaches a course on strategic management and information technology and his research projects deal with the causes and effects of electronic integration and information technology strategy, especially the management of infonnation technology outsourcing. His doctoral thesis was awarded the 1986 AT Kearney Award for Outstanding Research in General Management by the Academy of Management and his research papers have appeared in the Academy of Management Journal, Academy of Management Review. Information Systems Research, Journal ofManagement Information Systems, Management Science, Strategic Management Journaly and he has presented papers at several professional meetings. He serves on the editorial boards of Academy of Management Review. Journal of Management Information Systems, Organization Science, and Strategic Management Journal. This paper develops and tests a mode! of the determinants of information technology (IT) outsourcing by integrating both business and IT perspectives. Specifically, we attempt to explain the degree of IT outsourcing using business and IT competences as represented by their cost structures and economic performances. In addition, we posit that outsourcing is dependent on business govemance, particularly financial leverage. Based on factor analyses and multiple regressions using data from fifty-five major U.S. corporations, we observed that the degree of IT outsourcing is positively related to both business and IT cost structures. We also established that the degree of IT outsourcing is negatively related to IT performance. Finally, we conclude with implications and future research directions.
ABSTRACT.

Acknowledgments: Authors are listed alphabetically.Thisproject was supported by the research consortium on Managing Information Technology in lhe Next Era (MITNE) of lhe Center for
Information Systems Research (CISR) at the Massachusetts Institute of Technology. We tliank John Rockart. Judith Quillard, three journal reviewers, and the Editor-in-Chief for comments on earlier versions of the paper.
Journal o/MaMogemginfii/bmaiioHSysttms/SwrmusT 1992. Vol. 9. No, 1, pp. 7-24 Copyri^l 1992. M.E Sharpe. Inc.

LOH A>a3 VENKATRAMAN

KEY WORDS AND PHRASES: information

technology outsourcing information technol-

ogy strategy. (IT) has transcended its established administrative support function and has moved toward playing a more central role of business operations (see, for instance, [14,18,25,35]). Within this tradition, research efforts have focused on the use of IT to influence ihe boundaries of a firm with its suppliers, buyers, and other intermediaries [6,8,21]. There is, however, a glaring lack of research emphasis on the role of IT infrastructure as a component of the firm boundary itself. In other words, while IT has been considered a critical mechanism of multiorganizational business relationships, the research stream has treated the governance of IT infrastructure as if it were within one single firm's hierarchy. Such an approach fails to recognize the recent trend toward managing a firm's IT infrastructure through a variety of governance mechanisms with otber firms.
IT IS A TRUISM THAT INFORMATION TECHNOLOGY

Under contractual arrangements popularly termed "outsourcing," firms are increasingly shifting specific components of their IT infrastructure away from a "hierarchical" mode toward a "market" mode of governance. The well-publicized decision by Eastman Kodak to hand over its entire data center to IBM, its microcomputer operations to Businessland, and its telecommunications and data networks to Digital Equipment Corporation and IBM is a classic illustration [43,45]. Beside this particular case, it appears that IT outsourcing is becoming a serious strategic option for many firms. The Yankee Group estimated that all Fortune 500 firms would evaluate outsourcing, that a fifth of them would sign outsourcing deals during the 1990s, and that the outsourcing market would increase from $29 billion in 1990 to $49.5 billion in 1994 [7]. IT outsourcing can bo dependent on several factors across multiple levels. At the level of the economy, ilie temporal effects of trends and cycles may motivate firms to rationalize the management of the IT infrastnicture through arrangements like outsourcing. At an industry level, competitive pressures may induce firms to establish "partnership-based" relationships with key IT vendors. At the firm level, the quest for competitive advantage may serve as a critical impetus to the IT outsourcing decision. Within the firm, the decision to outsource may be dependent on several managerial factors. For instance, managers may like to build empires by accumulating control over corporate resources [27] such as the IT infrastructure. Further, the association of information with power [17] may inhibit the outsourcing decision. The practice of IT ou^wurcing has been extensively documented in the business periodicals, but there is scant attention provided to articulate its determinants. In otber words, we know the phenomenon in some detail but we do not fully grasp the set of factors leading to the outsourcing decision. Our objective in this paper is to develop a research model on the determinants of IT outsourcing. We recognize Ihc complex array of factors discussed above; but as a first attempt at establishing an empirical model, we focus on factors at \ht,firm level with appropriate controls for industry sector effects.

DETERMINANTS OF IT OUTSOURCING

Frameworks of IT Outsourcing
OUTSOURCING CAN BE FRAMED as a

"make-versus-buy" decision facing a firm. In its generic form, it has been studied in several settings, such as the manufacturing of parts in the automobile industry [30, 40], the sales function in the electronic industry [1], the procurement of components or services in the naval shipbuilding industry [23], and the distribution of equipment, componcnis, and supplies across a broad set of industrial firms 116]. Within lhe IT profession, lhe term "outsourcing" is often viewed as a buzzword that is confusing and often misunderstood [44]. We define IT outsourcing as the significant contribution by external vendors in the physical and/or human resources associated with the entire or specific components of the IT infrastructure in the user organization. This definition is consistent with the conceptualization of the IT infrastructure in terms of "the intemal organization of people and resources devoted to computer-based systems . . . [involving] both the tangible equipment, staff and applications and the intangible organization, methods and policies by which the organization maintains its ability to provide system services" [22, p. 148]. In the context of IT sourcing, vendors may contribute computer assets for the user. Alternatively, the ownership of certain computer assets of the user may be transferred to the vendor. Similarly, vendors may utilize Iheir personnel to provide the required services, or existing staff of the user may he employed by the vendor. In figure 1 we depict the distinction between outsourcing and insourcing based on the two dimensions central to our definition: (1) the degree of intemalization of physical resources by the user, and (2) the degree of intemal ization of human resources by lhe user. By intemalization, we mean the ownership of the computer assets or the employment of the system personnel. Several modes of the IT infrastructure have heen commonly outsourced by firms. These include applications development, data center, systems integration, systems design/planning, telecommunications/network, and timesharing. The modes of IT outsourcing vary through the different levels of contribution of physical and human resources by the user and the vendor. We also illustrate in this figure the typical location of each mode of outsourcing in the definitional framework. The various modes of IT outsourcing also differ in the domain of influence within the corporation. Domain of influence refers to the extent to which IT is inherent in the business processes as well as the administrative and functional coordination of the organization. For instance, an application development outsourcing arrangement ordinarily affects a specific domain of the firm, while a telecommunications/nelwork outsourcing arrangement may affect a more general domain of the firm. Further, an outsourcing arrangement differs in terms of the contractual mode (i.e., the type of relationship between lhe user and the vendor as governed by the agreement). For example, a systems design/planning outsourcing contract may be project-based, while a data center operations outsourcing contract may be period-based. In figure 2 we show the characteristics framework and illustrate the various modes of IT outsourcing along the two above dimensions.

10

LOH AND VENKATRAMAN

H.gh

INSOURCING

Intemalization of Human Resources

OUTSOURCING

I
I Low Low

Data
Center

I
I

I Application I Development

Intemalization ot' Physical Resources

H.gh

Figure L Outsourdng versus Tnsourcing

The Research Model

IN THIS PAPER, WE DEVELOP A MODEL of ihc determinants of IT outsourcing, with a particular emphasis on economic constructs from both the business and the IT contexls. Our framework is based on an argument that management should constantly reassess the scope of those activities that should be carried out with a firm's hierarchy and those that are best performed by extemal partners, including IT vendors. The guiding considerations for such make-versus-buy decisions include: relative cost advantage, and economies of scale and scope. Thus, it is likely that in some cases, an IT vendor may be a more appropriate entity to manage Ihe firm's IT infrastructure than the firm itself. This is because an outsourcing vendorbeing a specialistserves multiple users simultaneously. To the extent that the knowledge, skills, and capacity can be pooled across different customers, there ean be benefits of economies of scale, which are otherwise absent when single users perform the same tasks. In addition, the wide variety of IT projects undertaken by the vendors permits the reaping of economies of scope. The lower costs attributable to the vendor also arise from the enhanced IT competence and experience, both of which are absolutely crucial in managing the IT infrastructure in the complex and rapidly changing information era.

DETERMINANTS OF rr OUTSOURCING

General

Syslcms rtei Plan II ing

Telecoms/ Network

Center

Corporate Domain

Systems Integration

Data Process jpg Application Development

Specific

Project-Based Contraclual Mode Figure 2. CharacterisUcs of Different IT Oulsourcing Modes

Period-Based

We follow Henderson and Venkalraman's [12] model of aligning business and IT domains to derive ihe specific constructs of the research model. Thus, business and IT strategies are viewed as involving the dimensions of competence and governance. Further, we posit that IT governance (specifically, outsourcing) is dependent on the structural characteristics of the user organization, especially business competence, business governance, and IT competence. We elaborate our rationale below.

Business Competence
Business Cost Structure A well-accepted axiom in the strategy and economics literature is that a firm *s business cost structure (the entire spectrum of costs directly associated with tbe actual production and coordination of the firm's product line) is a significant source of business competence given its role in explaining business profitability (see, for instance, [5, 32]). Thus, firms try to produce their output below the average cost and are constantly under pressure in a competitive marketplace to reduce the relative cost of business operations. Given the ubiquitous nature of IT, which pervades the entire process of

12

LOH AND VENKATRAMAN

transforming inputs into outputs [33], the costs associated with a particular IT govemance include the direct technology cost and the indirect cost of supporting the administration of the enterprise. Thus, a firm in a situation of high relative cost will seriously consider the available options to reduce its business cost structure, including reassessing the positioning of its IT infrastructure within the scope of the firm's hierarchy. Therefore we hypothesize that a firm's business cost structure is a crucial determinant of IT outsourcing; Hypothesis 1: The firm's business cost structure will be positively related to the degree of IT outsourcing.

Business Performance Another component of business competence is reflected by the level of business performance. As noted in a trade periodical: "Reduced profits . . . are causing management to look everywhere to increase margins" [9,p. 89]. Under conditions of poor business performance, firms often seek to streamline their operations, including selling off or redeploying assets [10]. The traditional view of IT operations as an investment center or a service center is rapidly giving way to an emergent notion of a profit center. Thus, the IT infrastructure is no longer off limits to the top management team seeking superior performance. In fact, "much of what is fanning the fire f o r . . . outsourcing is that business is having to restructure to remain competitive" [9, p. 90]. When the firm does not perform well vis-5-vis its competition, the need to reevaluate the traditional govemance modes of all its major spheres of operations, including the IT arena becomes even greater. We thus seek to test: Hypothesis 2: The firm's business performance will be negatively related to the degree of IT outsourcing.

Business Govemance
Financial Leverage The need to reduce reliance on debt financing has been one of the key impetus to outsource the IT infrastructure. Indeed, as widely cited among practitioners, increased debt "has been a major reason for cutting costs in the IS area, thus supporting the use of outsourcing" [9, p. 90]. Within the context of an imperfect corporate financing environment (see [28]), financial leverage can result in problems relating to financial distress or bankruptcy [3] as well as agency [15]. Further, the cost of equity capital increases with financial leverage [13]. Debt and equity have been argued to be more than altemative financial instmments: they are different business govemance stmctures [47]. Accordingly, it is posited that the choice between debt and equity depends on the characteristics of the assets in

DETERMINANTS OF IT OUTSOURCING

13

which the funds are used. Debt govemance is more appropriate for financing redeployable assets, while equity govemance is more suitable for nonredeployable assets. Due to the complex and customized nature of systems, applicalions, and staff, the degree of redeployability of an installed IT infrastracture may be limited. Thus, debt governance is not the optimal foim of business govemance. A high level of debt hence results in a need to reduce the non-redeployable assets which then gives rise to a greater level of IT outsourcing. Thus we test Hypothesis 3: The firm'sfinancial leverage will be positively related to the degree of IT outsourcing.

IT Competence r r Cost Structure Investment in IT has recently escalated, and its importance is nowhere less evident than in its dramatic increase from $55 billion to $190 billion in the economy; in fact, IT accounts for about balf of most large firms' capital expenditures [18]. Due to the enormous outlay associated with the IT infrastructure, firms have found it necessary to adopt a better cost control approach to IT. In line with this notion, IT must be treated as a capital investment and not just an overhead of lhe firm. Firms have been plagued by the astronomic rise of IT expenditure in many specific IT areas that are necessary to run the business. For instance, in the area of application development, a critical problem has been the control of the cost of internally conceived software [11]. Consequently, corporations arc rationalizing iheir capital outlay on IT. Where possible, drastic restructuring of the traditional inhouse mode of IT govemance is undertaken to trim the high costs of IT infrastniciurc. As Weizerand Associates [42] put it: "Outsourcing can free capital tied up in data center hardware and save operating costs." An extremely attractive option available to firms is to outsource their IT infrastructure to value-added vendors who are more efficient in terms of managing and operating the IT. In three often-cited early cases of IT outsourcing, American Standard reportedly saved $2 million per year for its financial and payroll operations, Copperweld cut its systems budget from $8 million to $4 million, and Foodmaker slashed its data processing costs by 17 percent [36]. Other recent cases are Wabco and American Uliramar, which trimmed iheir annual processing costs from $3 million to S1.8 million, and from $3 million to $1.5 million respectively [46]. We thus seek to verify: Hypothesis 4: The firm's IT cost structure will be positively related to the degree of IT outsourcing. IT Performance With the elevation ofthe role of ITIrom the "backroom" to lhe "frontline" of business operations, firms are making IT directly accountable for its direct contribution to lhe

14

LOH AND VENKATRAMAN

overall corporate profitability. The profit-oriented posture imposed on the IT infrastmcture puts intense pressure on the technology to result in tangible economic returns. With the escalating level of IT investments needed to support business in the contemporary marketplace, there is a need to reconfigure the IT infrastructure in ways that make it possible to ascertain the benefits in a clear manner [38]. As IT expenditure has risen rapidly over the last decade [41], it is not surprising that managers arc more stringent than ever before in assessing the productivity of their IT infrasmicture. Thus, when economic profits fall in relation to IT investments, management faces an immense need to reevaluate the roleof IT. As efficiency of organizing is tied intimately to the mode of govemance, it is natural that there is a greater shift from the usual inhouse management to extemal involvement. Indeed, it has been the view within the practicing and consulting IT communities that "outsourcing is a key strategy that enable [companies t o ] . . . improve retum on equity" [31]. Therefore, we test: Hypothesis 5: The firm's IT performance will be negatively related to the degree of IT outsourcing. Figure 3 is a schematic representation of the research model with the five hypotheses.

Methods Sample
W E BEGAN WITH A SAMPLE FROM THE U S T of companies in a study of 200 major U.S. corporations carried out by G2 Research, Inc., which provided data on their level of IT outsourcing [24]. We also required that the level of total IT expenditure be available for operationalizing some of our independent constructs. Data availability across these two variables limited our study sample to fifty-seven firms (see the appendix for the list of companies). We collected corresponding data on the independent constructs pertaining to each firm in our sample for the fiscal year 1989 from Standard and Poor's Compustat II and Lotus' CD/Corporate on CD-ROM.' Thus, our data represents the use of both primary and secondary sources. During discussions with the managers of G2 Research, we ascertained that the data on outsourcing expenditure are an integral part of their professional service to their clients. Our overall assessment is that their method of data collection and verification meets the standards of our research purposes. Further, the integrity of the secondary data sources for the independent constmcts is widely accepted within accounting, economics, and finance research.

Operationali zation We need to operationalize four key constructs. For the degree of outsourcing (Y), we developed a ratio of IT outsourcing expenditure to total assets for each firm so that

DETERMINANTS OF FT OUTSOURCING

15

Business Governance

f 1

Financial Leverage

\ 1

Compelence H3(+) / ^ \ / Business \ 1 Cost Structure K

IT Competence

4 ^^ I I ^^ IT \^ Outsourcing 1

CosI Slructuro

V
f (

y
\ Y H2(-)

V
II5(-) 1 T

y
\ 1

Kzzy

Business Perfurmance

IT Performance

IT Governance

Figure 3. The Research Framework the level of IT outsourcing is normalized by firm size. The business cost structure (X{) was computed as lhe sum of the cost of goods sold and the selling, general, and administrative expenses, divided by net sales and total assets.^ Business performance (X^) was capttu^ed by retum of assets and earnings per share (fully diluted and excluding extraordinary items), representing the economic efficiency of lhe entire business as measured by its assets or equity. For financing leverage (X^), we look the ratios of long-term debt as well as of total liabilities with shareholder equity. Operationalization of constructs in the IT context is more difficult given the paucity of prior research. The metric used lo evaluate lhe effectiveness of IT has evolved from code analytic (1970-78), through design analytic (1978-84) and function analytic (1984-90), to thecurrent business directed (1990-present) measures [37]. In short, the criteria used are shifting from metrics thai focus on IT output to those that focus on the economic outcomes of IT activities. Further, it has been argued that the economic benefit of information systems can be best evaluated by measurements pertaining to the company's entire IT expenditure level, as opposed to the economic benefits derived from individual systems [4]. Thus, for IT cost structure (X4), we used (he ratio of IT expenditure with both gross plant, property, and equipment (i.e., before depreciation) and net plant, property, and equipment (i.e., after depreciation), which is analogous to the extent to which the physical infraslructure of lhe firm is represented by the IT infrastructure. IT performance (X5) was measured by ncl income and sales divided by IT expenditure, which corresponds to lhe economic efficiency of the IT assets.

16

LOH AND VENKATRAMAN

In addition, we specified two control variables, one for business size (X^) and another for industry (X^). The size variables included net sales and total assets, while binary dummy variables werc formed for service and industrial sectors.

Analysis
Due to the possibility of multicollinearity among our independent measures, we perfonned a factor analysis using the principal components method and a varimax rotation to discem the factor pattem inherent in the data structure. This procedure ensures thai the independent constructs used for our subsequent regression are orthogonal. We obtained the corresponding scores associated with a prespecificd set of the four factors for the business context and two factors for the IT context. In our mul tiple regression, we used the ratio of IT outsourcing expenditure to total assets as the dependent variable and the set of factor scores for the business and the IT contexts and the industry sector dummy variable as the independent variables. The econometric specification for our analysis is as follows: (1) r = po + piXi -

where e represents the random error.

Results Descriptive Statistics and Factor Structure

I SUMMARIZES THE MEANS AND STANDARD DEVUTIONS for all the individual indicators, while Table 2 provides the matrix of zero-order correlations. In Table 3, we depict the results of factor analysis for the business and IT contexts. With a varimax rotation, four factors in the business context can be interpreted: (1) business performance; (2) business cost stmcture; (3) business size; and (4) financing leverage. Two factors extracted for the IT context can be interpreted as: (1) IT cost stmcture, and (2) IT performance.
TABLE

Test of Hypotheses
The results of estimating equalion (1) are shown in Table 4. The overall model has an F-value of 2.10 (p < 0.06), and explains about 24 percent of the variance, which is acceptable for an exploratory research and a limited number of independent variables. In the business context, we accept HI since the coefficient for business cost structure is significant at the 0.05 level with the expected sign. In the IT context, we accept H4 and H5 since the coefficients for IT cost stmcture and IT performance are significant at the 0.1 and 0.05 levels, respectively. In contrast, the other two hypotheses (H2 and

DETERMINANTS OF IT OUTSOURCING

17

Table 1

Means and Slandard Deviations

Variable OSTA CSSA CSTA LDSE TLSE REOA EAPS ITGP ITNP NUT SAIT ASSE SALE SERV Mean 0.002578 0.8336 0.6754 0.4783 5.138 0.02667 3.089 0.1355 0.1522 2.385 56.91 21000 9405 0.5965 0.4035 Std Dev 0.004085 0.1555 0.6706 1.989 7.005 0.06765 5.257 0.1748 0.1738 7.298 44.81 26500 9877 0.4950 0.4950

Description Outsourcing Expenditure/ Total Assets Cost of Goods Sold-nSelling, General, & Administrative Expenses/ Sales Cost of Goods Sold+Selling, General, & Administrative Expenses/ Total Assets Lons-Term Debt/ Shareholders' Equitv Total Liabilities/ Shareholders' Equitv Return on Assets Earnings per Share ($) IT Expenditure/ Gross Plant, Property, & Equipment IT Expenditure/ Net Plant, Property, & Equipment Net Income/ IT Expenditure Sales/ IT Expenditure Total Assets ($ million) Sales ($ million) Service Sector Industrial Sector

INDU

H3) relating to business performance and financial leverage in the business context are not supported. Further, the two control variables did not emerge as significant.

Discussion
for our research model. First, HI was accepted, which suggests that the business cost structure is a critical determinant of IT outsourcing. A high level of business cost structure may motivate a firm to review its overall cost structure reflected in its costs of physical infrastructure (such as plant and equipment), including its IT infrastructure. Such restructuring may signal to the capital markets the strong commitment of corporate management to improve its business efficiency. Further, as IT pervades the entire value chain of the business, the adoption of IT outsourcing may result in a superior control of the business through the
THE EMPIRICAL RESULTS PROVIDE GENERAL SUPPORT

18

UJ

Ir
-

SAL

(N

c d

UJ

in

in

SI
<

rt;

91^
0220

1607

9
(N

d
QC

d
fNI

m 2307

(N

649

0.0831

p.

o
ON

fN .

c
^r
oc

d C o *^
in
IT)

c
C^
00 ""
QC

dL
tn
< o
cc
(N

d
IV
J ' -

d p
in

[^ . d I.^ 9 o

2499

o C:

o c

c c

o 499

279

)220

088

0.021

UJ

L.-

oc
J'

9
1010 0861 EOA

s d o
1375 o .1450

ri dC O c
in

s
0161

9
1161
d

gp
d

w o

0474

.2148

w en

(N

o- *~" 802

ir
dl.
(N C

0586

in ^
:

H 775
Ul

V
er
' .

9 t^
196
QO

o
ON

9 L'

108

p * in * d w
<:
OO

?
rN

c
in in

cp

o
tv

p
rN

(N

-o.r

-0.0,

-0.0

Ul

9 L-0.1599 -0.20
in in

i *

9 ^ 9 tl
-0.0120

d -0.20

to *

?(N

^l
-0.1586

d*

-0.0978 0.182

-O.05

-0.2769 -0.15

-0.0828 -0.09

in *

-0.01

CSS

I
o U o

*f

rN

a. o d -0.0411

tv

J 5
(N 9
tr

(N

p w

9 i-

L
IT.

OSTA

0.0386

ITi

0.0200

0.1671

O C

cc _ ^ OC
fN

r-i *

< in

F1/3

UJ U3

UJ

in

.^
UJ

(J

UJ

H
Ul

UJ <

>
UJ

=:

664

zoo
88800

DETERMINANTS OF IT OUTSOURCING

19

Table 3

Factor Analysis Results

3(a) Factor Structure for Business Context Factor 1 Eigenvalue Cum. Prop, of Var. Explained Variables CSSA CSTA LOSE TLSE REOA EAPS ASSE SALE
1 Business Pertormancel :.3?hl I) 2945

Factor 2
(Business Cost Slructure) 1.48bO

Factor 3
<Busines5 Size)

Factor 4
(FinancUi Leverage)

0.5428

1.3070 0.7062

0.S308 0.0187 -0.1233 0.9172 0.7561 -0.1503 -0.0324 0.2159 -0.0132

-0.0287 -0.0191 0.0270 -0.2323 0.9257 0.9257 -0.0586 0.1780

Rotated Factor Pattern 0,7909 -0.0194 0.S6U 0.0981 0.0513 0.4195 0.1128 0.0009 -0.0861 0.1319 0.0474 0.S2I7 -0.4225 0.1413 0.9.^21

3(b) Factor Structure for IT Context Factor 1

(IT Cost Struclurel

Factor 2
IIT Performancel

Einenvalue Cum. Prop, of Var. Explained Variables ITGP ITNP NIIT SAIT

2.0782 0.51%

1.0822 0.7902

Rotated Factor Pattern 0.97SS -0.0241 -0.0380 0.9814 0.S246 0.2032 0.6409 -0.3246

Note; See Table 1 tor a description of the variables.

"variable costing" of IT outsourcing, as opposed to the traditional "fixed costing" of inhouse govemance. Second, H4, which posited a positive relationship between IT cost structure and outsourcing, was empirically supported. As noted in a trade periodical, "Outsourcing is being seriously considered in more and more organizations as a poteniial solution to rising IS [information systems] costs" [9, p. 89]. To compete effectively in lhe modem informalion era, complex and costly systems are required to support and propel a corporation in its quest for competitive advantage. In particular, the key compelling force driving companies to outsource is cost savings: "In some cases, outsourcing vendors have promised to reduce annual IS outlays by 50%, although 15% to 30% savings are more common" [7, p. 47]. Thus, our empirical finding is consistent with the prevailing view regarding the need lo rationalize the IT cost structure in order to slay competitive in the market. Third, H5, which specified a negative relationship between IT performance and outsourcing, was also empirically supported. Low economic retums on IT investment appear to affect the propensity of firms to outsource more of their IT infrastructure to vendors. The present dilemma facing many IT executives appears to be a justification

20

LOH AND VENKATRAMAN

Table 4

Multiple Regression Results

OLS Estimate Standard Deviation 0.000902 0.000637 0.000623 0.000579 0,000621 0.000655 0.000543 0.001251
R2

Significance

Constant Busine.ss CosI Structure <Xi) Business Performance (X2) Financial Leverage (X3) IT Cost Structure (X4) IT Performance (X5) Business Size (X6) Service Sector (X7) Model Fit

0.{X)32I7 0.001498 0.000548 -0.000313 0.000806 -0.001520 0.000255 -0001017 F = 2.I0

()

()

= 0.24

Nole: (*"), ("), and (") denote one-tailed significiince at 0.01, 0.05, and 0.10 levels respeclively.

of investing in an IT infrastructure based on its productivity. Although there are several possible metrics to gauge the perfomiance of IT such as reliability, quality, timeliness, and the like, our results suggest that economic measures are valid indices to evaluate the productivity of IT in terms of the adoption of an efficient mode of IT govemance. Fourth, we did not find empirical support for H2which specified a negative relationship between business pcrfoimance and outsourcing and H3which specified a positive relationship between financial leverage and outsourcing. The general implication from these two results is that business performance and financial leverage may be too far removed in terms of influencing the level of IT outsourcing direcUy. Although we specified direct effects, which were not empirically supported, we urge that future research consider indirect effects (for example, through business and/or IT cost structures) or moderator effects. Finally, both business size and industry sector as control variables did not emerge as significant determ inants. The implication is lhat our results are generally robust and valid across firms differing in sizes (within the spectrum covered by the sample) as well as industry versus service categorization.

Toward a Comprehensive Model of IT Govemance

While we have provided some empirical support for a set of conventional wisdom regarding IT outsourcing in this paper, we recognize the limited scope of our model. We suggest a set of directions for refining this line of inquiry in lhe future. As we move away from the consideration of a firm-level focus using a neoclassical economic perspective, we identify three avenues of future research: (a) an organizationiU

DETERMINANTS OF IT OUTSOURCING

21

economic model of IT oulsourcing; (b) a diffusion process model of IT outsourcing; and (c) an organizational process model of IT outsourcing.

An Organizational Economic Model of IT Outsourcing Research in the general arena of make-versus-buy (including outsourcing) has been anchored within an organizational econ(Mnic perspective [2], especially transaction costtheory(forareview,see [48]). In addition.agency theory with particular emphasis on constructs such as goal alignment, incentive payment, and monitcning [15] has the potential to provide insights on the govemance of IT outsourcing. Moreover, it may be appropriate to reflect refinements such as the articulation of bargaining and influence costs [26] in the govemance of IT infrastmcture. For instance, constmcts relevant to bargaining costs may involve cowdination failure and information acquisition/asymmetry, and those related to influence costs may include exchange facilitation and competition foreclosure. While the present dataset did not allow us to operationalize ccmstmctsfroman wganizational economic perspective, a research design involving primary data from organizational informants would allow us to better understand and predict not only the degree of IT outsourcing (as done in the present smdy), but also the mode of outsourcing (see figures 1 and 2). We are inthe midst of such a study.

A Diffusion Process Model of IT Outsourcing In this paper, we have developed a "variance" model [29] for explaining IT outsourcing. We can adopt the view that an IT outsourcing arrangement constitutes an administrative innovation [19, 39]. Such an argument is based on the emergent departure of many firms firom an established hierarchical mode of goveming the IT infrastmcture toward a market mode. Outsourcing fundamentally transfomis the traditional requirements of managing IT from those rooted on an adversarial, "armslength" approach to those structured on a cooperative 'partnership-based' relationship. Using a macro-organizational level of analysis, the diffusion model will specifically examine the underlying force that motivates firms to adopt IT outsourcing. With this approach we can analyze whether the diffusion of IT outsourcing can be explained by imitative behavior. The findings are interesting in view of the prominence of the Kodak-IBM outsourcing contract which purportedly encouraged many other firms to consider such a govemance option seriously [34], We arc currently pursuing this diffusion-of-innovation study of IT outsourcing.

An Organizational Process Model of IT Outsourcing Another complementary extension would be to focus on the organizational routines underlying the management of this IT mode of govemance by developing an organi-

22

LOH AND VENKATTIAMAN

zational process model. Such a research initiative would go a long way in enhancing our understanding of the new phenomenon in the marketplace. The process model may incorporate new structures (e.g., shared authority, responsibility, property rights, and risk-bearing), management processes (e.g., allocation and coordination of resources, perfonnance assessment, and joint planning), and managerial roles (e.g., liaison, decision making, and leadership) that are required to derive the benefits from this mechanism effectively.

Conclusion
U.S. CORPORATIONS, we empirically identified a set of important determinantsreflecting both IT and business contextsof IT outsourcing. Thus, we offer the fu^t empirical assessment of a set of widely held assertions and bel iefs as to why firms outsource their IT infrastructure. We have also identified some important directions for extending ihis line of inquiryreflecting an organizational economic view, a diffusion-of-innovation view, and an organizational process view. We hope that our attempt at empirically testing this emergent phenomenon will stimulate others to look at this important strategic challenge facing firms from a rigorous theoretical perspective. Such research initiatives will allow us not only to better understand this complex phenomenon, but also to derive useful management prescriptions grounded on systematic theory-based research.
BASED ON DATA FROM LARGE

NOTES
1. The final sample size is 55 due to missing data in Compustat and CDAI^oiporale. Our sample comprises a rough split between industrial and service firms (about 4O--60). The selection is limited by the availability of data across multiple sources. It is possible that the primary data regarding IT expenditures may have been skewed, since our sources surveyed only large corporate users of FT (e.g., firms wilhin the Fortune 500 and Fortune Service 500 categories). In our sample, the level of IT expenditures as a percentage of revenue is 1.8 percent. This is compared with another survey [20] that obtained figures for ten different industries, ranging from 0.9 percent to 4.2 percent. Although the set of companies used is not a random sample of the entire population of firms in the economy, we believe that our choice of data sources allows us to include a sample that is representative of the set of major users of IT (i.e., large firms). 2. It is necessary to sum these two "costs" as the accounting treatment is not consistent for industrial and service fiims in the sample.

REFERENCES
1. Anderson, E. The salesperson as outside agent or employee: a transaction cost analysis. Marketing Science. 4.3 (1985), 234-253. 2. Bamey, J.B., and Ouchi, W.G., eds. Organizational Economics. San Francisco: JosseyBass, 1986. 3. Baxter, N. Leverage, risk ofruin and the cost ofcapital./oiwna/o/Financ^, 22,3 (1967), 395-403. 4. Bender, D.H. Financial impact of infonnation processing. Journal of Management Information Systems, 3, 2 (1986), 22-110. 5. Buzzell, R.D., and Gale, B.T. The PIMS Principles, New York: Free Press, 1987.

DBTERNUNANTS OF IT OUTSOURCING

23

6. Cash, JJ., and Konsynski, B. IS redraws competitive boundaries. Harvard Business Review, 63, 2 (March-April 1985), 134-142. 7. Eckerson, W. Changing user needs driving outsourcing. Network World, July 2,1990., pp. 1.47. 8. Gurbaxani, V., and Whang, S. The impact of information systems on organizations and maikets.Communicationsof the ACM, 34,1 (January 1991), 59-73. 9. Hammersmith,A.G. Slaying thelS dragon withoutsourcery.CompMferworW, September 18, 1989. pp. 89-93. 10. Harrigan, K.R. Strategies for Declining Businesses. Lexington, MA: Lexington Books, 1980. 11. Helms, G.L., and Weiss. I.R. The cost of internally developed applications: analysis of problems and cost control methods. Journal ofManagemeni Iriformation Systems, 3,2 (1986), 5-32. 12. Henderson, J., and Venkatraman. N. Strategic aligruneni: a model for organizational transformation through information technology. In T. Kochan and M. Useem, eds.. Transforming Organizations. New York: Oxford University Press, 1992. 13. Hsia, C.C. Coherence of the modem theories of finance. Financial Review, 16,1 (1981), 59-82. 14. Ives, B., and Learmonth, G.P. The information system as a competitive weapon. Communications of the ACM, 21, 12 (December 1984). 1193-1201. 15. Jensen, M.. and Meckling, W.H. Theory of the firm: managerial behavior, agency costs, and ownership sirucmre. Journal of Financial Economics,3,3 (1976). 305-360. 16. John, G., and Weiiz. B.A. Forward integration into distribution: an empirical test of transaction cost analysis. 7<:rna/o/Z.^w, Economics, and Organization, 4, 2(1988), 337-355. 17. Kanter, R.M. Power failure in management circuits. Harvard Bu.sine.ss Review. 57. 4 (July-August 1979). 65-75. 18. Keen. P.G.W. Shaping the Future: Business Design through Information Technology. Boston: Harvard Business School Press, 1991. 19. Loh. L . and Venkatraman. N. Infonnation technology outsourcing as an adminisEradve innovation: imitative behavior as an explanation of the diffusion pattern. MIT Sloan School of Management Working Paper. December 1991. 20. Maglitta, J.. andSullivan-Trainor. M.L. Do the right thing(s). Computerworld, September 30,1991, section 2. pp. 6-12. 21. Malonc. T.W.; Yates, J.; and Benjamin. R.I. Electronic markets and electronic hierarchies. Communications of the ACM, 30, 6 (June 1987). 484-497. 22. Markus. M.L. Systetns in Organizations. Marshfield, MA: Pitman, 1984. 23. Masten. S.E.; Meehan, J.W.; and Snyder, E.A. The costs of organization. Journal of Law. Economics and Organization, 1, 1 (1991). 1-25. 24. McCormick, JJ. Outsourcing action. Iriformation Week, September 10. 1990, pp. 84-92. 25. McFarlan, F.W. Infonnation technology changes the way you compete. Harvard Business Review. 62, 3 (May-June 1984). 98-103. 26. Miigrom. P.. and Roberts. D. J. Bargaining costs, influence costs, and the organization of economic activity. In J. Alte and K. Shepsle, eds.. Perspectives on Positive Poiiticai Economy. Cambridge: Cambridge University Press. 1990. 27. Mintzberg,H./'owc/-/na<i4ronJOrgarti'zfl/io?.EngIewoodCliffs,NJ: Prentice-Hall, 1983. 28. Modigliani, F., and Miller, M. Corporate income taxes and ihecost of capital: acorrection. American Economic Review, 53. 3 (1963), 433-443. 29. Mohr, L. Explaining Organizational Behavior. San Francisco: Jossey-Bass, 1982. 30. Monteverde, K.. and Teece. D. Supplier switching costs and vertical integration in the automobile industry. Be//yorrw/o/"conCTniCi', 13,1 (1982), 206-213. 31. Oltman. J.R. 21st century outsourcing. Compureru-or/rf. April 16,1990, pp. 77-79. 32. Porter, M.E. Competitive Strategy, New York: Free Press. 1980. 33. Porter. M.E., and Millar, V.E. How infonnation gives you competitive advantage. Harvard Business Review, 63.4 (July-August 1985), 149-160. 34. Radding, A. The dollars and sense of outsourcing: the ride is no bargain if you can't steer. Conyjuterworid, January 8,1990, pp. 67-72. 35. Rockart, J.F., and Scott Morton, M.S. Implications of changes in infonnation technology for corporate strategy,/n/er/acei. 14,1 (1984), 84-95.

24

LOH AND VENKATRAMAN

36. Rothfeder, J. More companies are chucking their computers. Business Week, June 19, 1989, pp. 72-74. 37. Rubin, H. Measure for measure. Computerworld, April 15, 1991, pp. 11IS. 38. Strassmann, P.A. The Business Value ofComputers. New Canaan, CT: The Information Economics Press, 1990. 39. Teece, D. The diffusion of an administrative innovation. Management Science, 26, 5 (1980),464--470. 40. Walker, G., and Weber, D. A transacdon cost approach to make-or-buy decisions. Administrative Science Quarterly, 29, 3 (1984), 373-391. 41. Weill. P., and Olson, M.H. Managing investment in information technology: mini-case examples and implications. MIS Quarterly, 13,1 (1989), 3-17. 42. Weizer, N., and Associates. The Arthur D. Little Forecast on Informaiion Technology and Productivity: Making the Integrated Enterprise Work. New York: John Wiley, 1991. 43. Wilder, C. Kodak hands processing over to IBM. Computerworld, July 31,1989, pp. 1,6. 44. Wilder, C. Outsourcing: fad or fantastic? Computerworld, December 25/January 1, 1989-90, p. 8. 45. WUder, C. DEC. IBM play ball in Kodak deal. Computerworld. January 15.1990, p. 8. 46. Wilder, C. Outsourcing: from fad to respectability. Computerworld, June 11,1990, pp. 1,122. 47. Williamson, O.E. Corporate finance and corporate govemance. Journal of Finance, 43, 3 (1988), 567-591. 48. Williamson, O.E. Transaction cost economics. In R. Schmalcnsee and R.D. Willig, eds.. Handbook of Industrial Organization. Amsterdam: North-Holland, 1989. APPENDIX: List of Companies in Aetna Life and Casualtv Co Air Products and Chemicals Amax Inc Armco'Inc Ashland Oil Inc Atlantic Richfield Co Bank ot Boston Corp Bankamenca Corp Baxter International Inc Champion International Corp Chrysler Corp Consolidated Rail Corp Continental Bank Corp Control Data Corp Corning Inc CSX Corp Dow Chemical Co E I Du Pont De Nemours and Co Duke Power Co Eastman Kodak Co Entergy Corp First Bank System Inc i Fleming Cos Inc Geico Corp General Dynamics Corp General Electric Co General Re Corp Great Western Financial Corp GTE Corp

the Sample
Harrier Inc Hometed Corp Illinois Power Co IngersoU Rand Co Keycorp Lincoln National Corp Mack Trucks Inc Manufacturers Hanover Corp NCR Corp Norfolk Southern Corp PNC Financial Corp Quantum Chemical Corp Reynolds Metals Co Rockwell International Corp Rohm and Haas Co SCECorp Security Pacific Corp Sun Co Inc Texaco Inc Textron Inc Travelers Corp UAL Corp Union Carbide Corp Unisys Corp United Technologies Corp U S Air Group Inc Valley National Corp Westinghouse Electric Corp