Dependent variables for the study of firm and industry-level impacts of information technology

(

HD28

.M414

^?

0/^,

-5. \»SI. U

^

NOV 12 1987

Center for Information Systems Research

Massachusetts Institute of Technology

Sloan School of Management

77 Massachusetts Avenue

Cambridge, Massachusetts, 02139

Jr

DEPENDENT VARIABLES FOR THE STUDY
OF FIRM AND INDUSTRY-LEVEL IMPACTS
OF INFORMATION TECHNOLOGY

J. Yannis Bakos

August 1987

CISRWPNo. 161
Sloan WP No. 1923-87

• 1987 J. Yannis Bakos

This paper will be presented at the Eigth International Conference on Information

Systems, Pittsburgh, PA, December 1987.

Center for Information Systems Research

Sloan School of Management
Massachusetts Institute or Technology

Dependent variables for the study of

firm and industry-level impacts

of Information Technology

J. Yannis Bakos

Center for Information Systems Research

Sloan School of Management

Massachusetts Instistute of Technology

Cambridge, Massachusetts

August 1987

ABSTRACT

Macro-level impacts of information technology (those at the level of entire
organizations, industries or the society as a whole) have not been studied in
the depth accorded to impacts at the individual user level. Furthermore,
there is a lack of studies that can claim to have successfuly demonstrated
specific impacts at this level. We believe that a well-defined and
instrumented set of dependent variables at this level would make a
significant contribution in this area. /^This paper addresses the issue of
identifying the appropriate dependent variables for research on the impacts
of information technology at the firm and industry levels. The role of
different organizational perspectives is examined, and some implications for
the design of empirical studies are discussed. |

Copyrighted 1987 by Yannis Bakos

INTRODUCTION

Studying the organizational impacts of information technology is at the core of
the information systems discipline, as evidenced by the substantial body of
exising literature on the nature of these impacts at different levels, ranging from
the individual user to the level of society as a whole. Ever since Mason and
Mitroff (1973) posed the MIS problem in a task-focused setting, impacts of
information technology at the individual level have provided the major focus of
MIS research. Several efforts have been made to establish a number of
theoretical propositions and build a cumulative tradition. One sign of the relative
maturity of research at the individual level is the dominance of empirical studies
and validated methodologies over work based on assertions and speculation.
Although room for improvement may still exist, research on the impacts of
information technology at the level of entire organizations, industries, or society
as a whole is still adolescent by comparison.

As any area of research matures, it becomes necessary to rigorously formulate its
theoretical propositions and subject them to empirical testing, thus creating a
need for the identification and measurement of the relevant dependent and
independent variables. Issues related to the characterization and operatio-
nalization of information technology* as the independent variable have been
addressed elsewhere; in this paper we focus on the identification of appropriate
dependent variables for the study of the organizational and industrial impacts of
information technology. Although an agreement on research variables is not a
necessary prerequisite for research, it would provide several benefits, such as a
more direct link to the underlying research disciplines, the possibility to transfer
operationalizations and measures across settings and to generalize and compare
the results of difTerent studies, and a start toward the building of a cumulative
tradition of research.

Research on user information satisfaction (UIS) at the individual user level can
provide an example demonstrating these benefits. Since UIS was established as a
generally accepted important variable at the individual level, there has been
significant improvement in our ability tu understand it, measure it, and design

proper studies that are more likely to produce rigorous results. Furthermore, a
lot of controversy was generated about the proper definition, role, importance,
use, and factors underlying or affecting this concept, which was constructive in
refining its theoretical underpinnings and clarifying its practical significance.

This paper reviews the major research approaches related to the impacts of
information technology at the organizational and industry levels, and argues that
at this point information systems research should build on, and advance, relevant
work in organization theory and industrial economic^. One of our primary goals
is to illustrate how the alternative perspectives on action and different
organizational views in the underlying disciplines of organization theory and
economics have different implications for the choice of the appropriate dependent
variables. We also address some theoretical and empirical problems in studying
the impact of information technology on these variables.^

ORGANIZATIONAL I'KKSPKCTI VKS AND IMPACTS RESEAKCH

In this section we present a meta-theoretical scheme for classifying the major
perspectives in the organizational literature, based on three criteria: the level of
organizational analysis, the dynamics of interaction among organizational units
at each level, and the underlying assumptions about human nature.

Levels of Analysis

Organizations can be viewed at different levels of analysis, and efforts to predict
or understand the impact of information technology are typically focused on one
of these levels. Most existing literature is consistent with the following five-level
view of organizations:

• an individual performing a task

• a work group including many individuals

• an organization consisting of several groups

• an industry with a number of firms

• the entire economy, or society as a whole.

i

Views on interaction

The hierarchical view presented above assumes that each organizational level
can be seen as a system composed of relatively independent subunits of the next
lower level, which may or may not differ in their goals and behavior. This leads to
two points of view: assuming goal congruence or coherent action among
organizational units at a certain level, we can focus on the structure and process
characteristics of organizations; alternatively, assuming divergence in the goals
and behavior of individual units, we can study the dynamics of their potentially
conflicting interactions and their effect on the behavior of the composite system
as a whole. These viewpoints, which we call structural and dynamic, will suggest
difTerent types of organizational impacts of information technology, and different
dependent variables.

Assumptions about human nature

Three major perspectives on action emerge in the organizational literature, based
on different assumptions about the nature of people and organizations:

• The rational (mechanistic) perspective assumes that action is purposive,
intentional, and goal directed. People and organizations have goals, and
they take actions to achieve them. This perspective underlies most work
based on mathematics, systems theory and economics. It has also been
dominant among information systems researchers.

• The externally constrained (evolutionary) perspective sees action as
environmentally driven. People and organizations react in response to
environmental constraints that leave little room for individual
rationality, their primary goal being survival perse.

• In the organismic (process) perspective, action is emergent, almost ran-
dom, dependent on organizational processes and social structures. Any
rationality is posterior, and goals often emerge as a by-product of the
process.

Organizational perspectives and the selection of dependent variables

We have identified five levels of organizational analysis, two views of organiza-
tional systems, and three perspectives on action. Figure 1 illustrates this scheme
for the organizational levels of interest in this paper, and provides examples of

structural

dynamic

Perspectives

rational

constrained

organismic

Levels

society

industry

organization

- ninrrorronomic y^

impact"! j/^
■ market failures/'^

■ population
ecology

/^

- ethnoloBiral ,''
appronchps ,'

■ inslitutionalizaufin
theories y'

y^ ■ Marvist

y^ class strueple

y^ pprspprtivps

y^

- rrsource

dependence

theories

•

-industrial y^
frnnoniics y^

(structure) y^

- strntrpic
alliances

/

^

•
•

yy - industrial

yy econniiiic?

^y Iconrlucti

^

- corporate
strategies

-structural y^
continRpncy y^

• information y^
processing y^
view yy

yy -political
^y Ihporips

- population
ecology

^y - coenitivp theories ,''
yy of breanizations ,'
y^ - institution8li2at4(5n
y^ theories ^^'
- sociotpchpnal
- resource Iheorie-
dependence - orgaitlzed anarchies
theories p«Cspective

Figure 1: Alternative organizational perspectives

major approaches falling into its subcategories. Ashley and Van de Ven (1983)
have proposed an alternative scheme, essentially based on the third and the first
of these dimensions: "(our scheme is based on) two sources of antithesis (which
are) manifested in structure-action and part-whole dialectics" (p. 245). We
strongly recommend to the interested reader their discussion of the underlying
debates on the nature and structuring of organizations, although they conclude
that these debates, representing conflicting world views, are unlikely to be
resolved in the foreseeable future. However,

"an awareness of the underlying values and biases upon which the theory is
constructed becomes essential. These values and biases act as assumptions, taken
for granted, in the world views that guide theorizing, and they constitute
paradigms that channel attention in specific directions and preclude the
investigation of alternative theoretical, ideological, and practical spheres " (p 270).

r

I

I

IT

(1)

(2)

Structure and
Processes

(3)

1

Performance

Figure 2: Areas for impacts research

In the context of organizational impacts of information technology, alternative
perspectives lead to different dependent variables and suggest the use of different
theoretical tools for the study of these impacts. Studies based on different
perspectives have used different vocabularies and, as a result, often have talked
past each other. A simple model for the impact of information technology is
shown in Figure 2. The technology has an impact on organizational structure and
process, thereby affecting organizational performance. This model defines three
areas for impacts research: (1) the impact of information technology on
organizational performance, (2) the impact of information technology on
organizational structure and processes, and, (3) the impact of organizational
process changes on organizational performance.

The general problem of the link between organizational structure, process and
performance variables is best left to organization theorists. Information systems
researchers may occasionally need to extend the available results in certain
areas, but as Keen (1980) has emphasized, such research is worthwhile only
when it makes a contribution to the underlying discipline of organization theory.
The majority of impacts research will belong to one of the first two areas: impact
of information technology on organizational performance and on organizational
structure and processes. The difference between the two areas can be visualized
as whether the structure and process box in Figure 2 is seen as a system that can
be modeled and probed, or as a "black box" whose inputs and outputs are the only
observable variables.

IDENTIFYING FIKM-LEVEL VARIABLES

In this section we review the literature on organization-level impacts of infor-
mation technology and use the alternative perspectives on action (rational,
constrained, and organismic) to identify relevant dependent variables.

Rational approaches

Research employing the rational perspective has focused either on organizational
performance or on structure and process variables. The major research problem is
to identify the impacts of information technology on organizational structure and
processes, and to determine how they can be predicted, controlled or mediated to
fit organizational goals. Two major views have emerged in the non-political
literature. Structural contingency theorists like Chandler (1962) postulate that
organizational structure affects organizational performance, and they attempt to
determine the nature of this relation. The information processing view focuses on
the information processing capacity of organizations and their effort to cope with
the constraints of individual bounded rationality and environmental uncertainty
(Galbraith 1973, Tushman and Nadler 1978).

Political models of organizations generally assume rational, intentional action
among organizational subunits, to achieve their divergent goals and objectives.
Coalitions of organizational subunits have been the main unit of analysis in
political approaches (Cyert and March 1963). Two central organizational
variables have been proposed: coalition structure, focusing on the process of their
formation among eligible participants (Pfeffer 1978, 1981), and coalition power,
most often defined in terms of the capacity to overcome opposition (Blau 1964,
Salancik and Pfeffer 1977), or as the ability to influence organizational goals
(Cyert and March 1963).

Environmentally constrained (evolutionary)approaches

Population ecology is the major approach employing this perspective from a
structural point of view. It focuses on the characteristics of environmental
selection (evolutionary change). Its underlying assumption is that organizations

are subject to environmental constraints to a much larger extent than the ratio-
nal view will concede, leaving little room for rational action (Freeman 1982).
Variables that have received attention are organizational type, such as specialist
us generalist; environmental characteristics, such as a fast or slow changing,
finely or coarsely grained environment; degree of competitiveness, characterized
by its institutional framework or by the resulting organizational mortality; and
organizational adaptability, such as learningcapacity.

In the dynamic view, resource dependency theories focus on managing external
dependencies. They are similar to rationalist political theories as they start from
a coalition model of organizations, but the emphasis is on the type and importance
of environmental dependencies. Power is assumed to accumulate to coalitions
managing these dependencies, which may arise from environmental constraints,
contingencies, resources, or uncertainties.

Organismic appioaches

The organismic perspective employs an emergent, almost random, process-
constrained view of action, and is not as homogeneous as the other two. There is
no clear distinction between structural and dynamic views, because conflict
among subunits is not a central concept. The view of organizations as organized
anarchies (Cohen, March, and Olsen 1972, March and Olsen 1976), stresses the
sequential, unfolding nature of organizational action. Behavior cannot be predi-
cted either by the preferences of individual actors or by environmental conditions.
Instead organizations are seen as "garbage cans" in which people, problems, solu-
tions and opportunities for action are mixed together, the results largely
depending on the process of interaction of the ingredients. Rationality and prefe-
rences are seen as emerging retrospectively from action, rather than guiding
action (March 1978).

Cognitive perspectives see organizations as providing paradigms (models of the
world), myths (perceptions of the world), and a common language that give
meaning to the life of individual participants (Weick 1979). These approaches
emphasize the importance of harmony between the technological and social
components of an organization (Trist 1981 ). or focus on quality of work life issues.

• structure

- goverance mechanism

- centralization/decentralization

- integration/differentiation
-coalition formation

- power distribution

- specialization, niche creation

• process

-efficiency

- production costs
-coordination costs

- flexibility

- information processing capacity

- political friction

- coalition structure

- coalition power

- adaptability

• performance

- economic returns

- survivability

Table 1: Organization-level variables

In this perspective information can be seen as a symbol rather than an
instrument (Feldman and March 1981).

Table 1 summarizes the major structural, process and performance variables
suggested by the rational and constrained perspectives, which are likely to be
affected by information technology. The organismic perspective is too fragmented
to allow us to abstract a concise set of variables, and although there is a
substantial body of literature looking at the links between organismic processes
and performance at the individual level, such as the relation between job
satisfaction and performance (Vroom 1964), these results have not been extended
to the organizational level. We are not trying to diminish, however, the signi-
ficance of the constraints of social and organismic processes for the design and
implementation of information systems, as several failed system implement-
ations could attest to their importance.

8

FIRM-LEVEL IMPACTS KESEAKCH

In this section we address the problem of identifying dependent variables for the
firm- level impacts of information technology. We look into three areas, organiza-
tional structure, organizational processes and organizational performance.

Organizational Structure

The impact of information technology on structural variables has received
significant attention in the literature. Kling (1980) surveyed several theoretical
and empirical studies of the relation between computers and organizational
structure, primarily in the period 1970-1979. Most early studies, however, used
information technology as the dependent variable, as they tried to explore the
factors influencing its adoption by organizations. Carter (1984) and Robey (1981)
conducted studies using organizational structure as the dependent variable.
They looked at authority relationships, centralization of control, differentiation,
and horizontal relationships among departments, with generally inconclusive
results. Attewell and Rule (1984) reviewed theoretical and empirical evidence on
the impacts of information technology on centralization, differentiation of organi-
zational subunits, and vertical and lateral communications. They concluded that
the evidence was fragmentary and mixed. Kriebel and Moore (1982) suggested
that the technology will affect the centralization of organizational structures and
the mechanisms for control and resource allocation.

Thus no direct link has been established between information technology and
organizational structure or performance, and the evidence from the organization
theory and strategy disciplines suggests that any such link would depend on the
particular organizational and environmental contingencies. This would explain
theconOicting results of Attewell and Rule (1984) and Robey (1981), who reported
that the technology is likely to either centralize or decentralize, differentiate or
integrate, and promote functional or less centralized hierarchies; a fact which
organizational theorists pointed out several years ago (Zannetos 1970).

Organizational Processes

Process variables combine the advantage of a direct theoretical link to organiza-
tional performance, and the potential for study at the level of organizational
subunits, subsequently aggregating the observed impacts. Unfortunately there is
little empirical research in this area to guide us in building and testing specific
theoretical hypotheses and in the operationalization of the underlying constructs.
The role of information technology in relieving constraints in the storage,
processing and communication of information caused by bounded rationality
(Bakos 1985), and in increasing organizational flexibility (Malone and Smith
1984), can provide the basis for the identification of organizational variables
susceptible to impacts from information technology.

Malone (1985) and Malone and Smith (1984) suggest that organizational
technology can be characterized in terms of production and coordination
technologies. Thus efficiency at the organizational level can be viewed as
dependent on production costs and coordination costs. Frequently this distinction
depends on the level of analysis employed, since what may appear as a production
cost at a certain level of analysis, can often be decomposed into production and
coordination costs at a finer level of detail. Since coordination costs come from
processing and communicating information, they provide a definite basis to
establish a theoretical link between information technology and efTiciency. The
accounting discipline has addressed the problem of measuring efficiency gains of
organizational subunits, but their approaches are geared toward the
measurement of production costs. Research on impacts of information technology
would greatly benefit from the systematic development of tools to help us assess
coordination costs, since they seem to be a primary beneficiary of the technology,
at least for some types of systems (Crawford 1982). Moving in this direction
Ciborra (1985) used transaction-cost theory to discuss the impact of the
technology on the costs of coordination and control, and on the organizational
governance mechanism.

Organizational flexibility provides a measure of long-term efficiency in a
changing environment. Malone and Smith (1984) defined it in terms of
vulnerability (the cost of expected failures) and adaptability (the cost of

10

adjustment). Williamson ( 1975) employed the related concept of asset specificity,
i.e., the best alternative use of an asset outside its present economic engagement.
We would prefer a definition comparing the expected efficiency in a stable
(although not necessarily deterministic) environment with the expected effici-
ency under varying degrees of environmental uncertainty. The performance of
flexible systems would be subject to less relative degradation as uncertainty
increases. Using results from queuing theory, Malone and Smith were able to
compare the relative efficiency and vulnerability of several simple organizational
forms. Similar approaches could be used to derive theoretical predictions for the
link between information technology and organizational fiexibility (for a rational
perspective) or adaptability (for an environmentally constrained perspective). In
relation to adaptability, Argyris (1982) has pointed out the potential of
information systems to reinforce different types of organizational learning.

The notion that information technology will increase the information processing
capacity of organizations has probably been an assumption for researchers in the
information systems discipline, and hence there exists no empirical research to
test it. Malone and Smith (1984) looked at the impact of information technology
on vertical and horizontal information fiows, the structure of organizational
networks, and alternative coordination mechanisms. Related measures could be
provided by looking at the quantity and quality of decisions or other outputs of
subunits, and reaction times to expected and unexpected contingencies.

Theories on the determinants of organizational power have proposed a number of
variables explaining the accumulation of power to certain organizational
subunits, like coping with critical uncertainties (Thompson 1967), or control of
information or other resources (Salancik and Pfeffer 1977). Game theoretic
approaches would suggest a focus on the goals and objectives of the individual
participants, the structure of the payoffs under different outcomes, and the costs
of coalition formation. Allocation of budgets, succession to administrative
positions, policy decision patterns, choice of information systems, have all been
studied as power-related dependent variables. In the MIS literature, Markus
(1983) has pointed out the potential of information systems to affect the
distribution of intraorganizational power, and has identified resource dependency
considerations as relevant for the implementation of information systems.

11

A number of authors have adopted an organismic perspective in looking at the
impacts of information technology. Kling (1980) reviews a number of articles
with a similar "segmented institutionalist" perspective. Bostrom and Heinen
(1977) focus on sociotechnical considerations in the design of information
systems. Markus (1983) looks at the interaction between the information system
and the organizational setting. Attewell and Rule (1984) cite controversial
evidence on the role of information technology in the humanization or
dehumanization of work.

Organizational Performance

Firm-level performance has been proposed as a dependent variable as well
(Crowston and Treacy 1986). A number of MIS researchers have used dependent
variables from the accounting literature for the assessment of economic
performance, such as the productivity of the firm or the information system per se
(Kriebel and Raviv 1982), return on assets and investment (Cron and Sobol 1983),
or cost/benefit assessments of the value of information systems (Crawford 1982).

It should be realized, however, that empirical inquiry into the direct link between
information technology and organizational performance is a particularly difficult
task: there are too many confounding factors and sources of extraneous variance,
necessitating accurate operationalizations of the independent variable, as well as
use of methodologies allowing for controlled settings, such as field experiments.
Carrying this sort of experimentation with entire organizations is next to im-
possible, therefore we must focus on the variables that will allow us to look at a
finer level of detail, i.e., organizational structure and process variables.

IDENTIFYING INDUS IRY-LEVKL VARIABLES

Industrial economics and corporate strategy provide the theoretical disciplines for
the study of the impacts of information technology at the industry level. The
notion of conflict coming from the divergence in the goals and objectives of
individual organizations, is central in this literature, and consequently this level
of analysis is dominated by the rational and environmentally constrained
perspectives, which pay attention to preferences, goals and objectives, and thus to

12

potential conflicts. We shall use the more popular terms "industrial structure"
and "industrial conduct" to refer to the strutural and dynamic views as applied to
the industry level.

Rational perspective

Rational approaches have focused on conflict among firms in the study of
industrial conduct, and the comparative dynamics of entire industries in the
study of industrial structure. Economic theory has provided the primary
reference discipline, and some measure of economic returns relative to the rest of
the industry or the rest of the economy is the prevalent indicator of performance.
The existence of an underlying analytic discipline has led to a focus on explana-
tory and causal, rather than descriptive, variables. On the other hand the
validity and relevance of these variables, the related approaches, and the under-
lying assumptions, have been often questioned (Kuttner 1985).

Three major concepts underlie the rational industry-level approaches: efficiency,
market power, and sustainability, which appear both at the industrial structure
and industrial conduct levels. Efficiency refers to the ability to maximize outputs
for a given set of inputs. Power refers to tht ability to resolve conflicting situa-
tions to one's advantage. Industrial economics has examined at length the link
between comparative efficiency, power and superior economic returns for
individual firms or entire industries. Sustainability refers to the ability to
preserve these superior returns over time. The balance of power at the
boundaries between entire industries is determined by the collective power of an
industry against its customers (monopoly power), its suppliers (monopsony
power), and the degree of collusion among industry participants (rivalry among
industry competitors). The sustainability of superior economic returns is
determined by the threat of substitution, and by the barriers to new entry, such
as the existance of potential entrants, the cost of entry, the likelihood and
credibility of retaliation by current competitors (Porter 1980).

At the industrial conduct level, comparative efficiency refers to the ability of a
firm to produce a product using fewer inputs relative to other products considered
as equivalent. Comparative efficiency can be achieved through improved
internal (intraorganizational) efficiency, or through the identification of

13

interorganizational synergies. Market power reflects the ability of a firm to
resolve conflictual situations against its customers and suppliers to its own
advantage. The major related variables proposed in the economic and strategic
literature are market share, power against suppliers, power against customers,
unique product features, positional advantages, and asset specificity (defined as
the best alternative use of a firm's assets). Variables that seem to determine the
ability of some firms to sustain superior economic returns relative to their
competitors include the capability for continued innovation, the ability to make
credible threats of retaliation against acts that might erode their competitive
advantage, and the quality and timing of their strategic moves. As discussed in
the next section, several of these variables could be affected by information
technology.

Constrained perspective

This perspective has been underlying much of the traditional strategic literature,
especially theories related to the identification of generic strategies. Population
ecology and resource dependency theories provide the major concepts. Their
theoretical imperative is that in the long run well focused, specialist strategies
outperform unfocused, generalist ones, and that firms strive for the control of
critical environmental resources in an effort to negotiate the uncertainty and
hostility of their environments.

At the industrial structure level, relevant variables that could be affected by
information technology include the rate of environmental change, net mortality
(births and deaths) of organizations, and changes in industrial boundaries to
reduce uncertainties coming from the technology, e.g., vertical information
integration. At the industrial conduct level, variables likely to be affected by
information technology include the ability to adopt major specialization
strategies, such as low cost leadership, differentiation, or focus on market niches,
the creation, restructuring or dissolution of organizations, and mergers,
acquisitions, or strategic alliances to control critical information dependencies
and to exploit opportunities created by information technology.

Table 2 shows the major variables in the literature used to characterize the
structure of industries which may be affected by information technology. Table 3

14

• efficiency

- procJuction economics

- economies of scale
-economies of scope

- economies of specialization

- internal efficiency

- production costs

- coordination costs

- interorganizational synergies

-coordination costs

- transaction costs

• market Structure

- market form

- market efficiency

- market power

- competitive rivalry

• sustainability

- barriers of entry

- threat of substitution

- credibility of retaliation

- rate of innovation

• survivability

- net mortality

- strategic alliances

- M&As, vertical information integration

• performance

- economic profits

- excess returns

Table 2: Industrial structure variables

shows the five major variables identified as determinants of a firm's conduct
within an industry. As discussed earlier, the former three come from rational
(economic) approaches, while the latter two are suggested by environmentally
constrained (evolutionary) perspectives.

INUUSTRY-LEVEL IMPACTS RESEARCH

Identifying the industry-level variables which are likely to be affected by
information technology, and understanding the nature of these impacts, is at
least as difficult as the corresponding problem at the organizational level.

15

• comparative efficiency

• market power

- market share

- unique product features

- customer and supplier switching costs

- asset specificity

- positional advantages

• sustainability

- product and process innovation

- threat and credibility of retaliation

- timing

• survival strategies

- niche creation

- differentiation

- cost advantage

• control of critical resources

- strategic alliances

- vertical information integration

Table 3: Industrial conduct variables

Industrial structure

The main focus of the traditional MIS literature has been on efTiciency-related
impacts of the technology. As Kriebel and Moore point out (1982), information
technology will affect industry-wide efficiency through its impacts on the
economics of production, most likely by changing the economies of scale, scope,
and specialization. Another type of efficiency-related impacts is likely to come
from interorganizational synergies, primarily through coordination efficiencies
(Cash and Konsynski 1985, Malone 1985) and reduced transaction costs (Ciborra
1985).

More recently, Ives and Learmonth (1984) and Beath and Ives (1986), used the
efTiciency of internal processes as their dependent variable. Parsons (1983)
pointed out the potential impact of the technology on competitive forces (buyers,
suppliers, substitution, new entrants, competitive rivalry) and industry-wide
economics of production. Cash and Konsynski (1985) elaborated on Parsons'
variables to suggest a more detailed set (capacity, cost structure, prices, entry

16

barriers, quality, services, competition), and Bakos and Treacy (1986) identified
search costs, unique product features, switching costs, bargaining power, internal
and interorganizational efficiency, as variables which could be affected by the
technology.

Furthermore, information technology can affect the monopoly or monopsony
power of an industry and the competitive rivalry among industry participants,
e.g., by requiring substantial non-recoverable investments, changing the costs of
exit, and altering the relation between fixed and variable costs. It can also affect
the sustainability of superior economic returns enjoyed by an industry with
market power. All four variables in Table 2 related to sustainability (barriers to
new entry, threat from substitute products, likelihood and credibility of
retaliatory action against new entrants, and rate of innovation in the industry)
are potential candidates. At this level of analysis the net effect of the technology
is uncertain and highly dependent on industry characteristics, and thus more
detailed variables will be necessary for a normative model of the relevant
impacts. Empirical reseearch beyond the case-study level is suspect, because of
confounding variables, extraneous variance, and difficulty in the
operationalization of the underlying variables.

From an evolutionary perspective, information technology is likely to affect
variables related to the ability of an industry to survive and evolve in a changing
environment, as well as the rate of environmental change faced by that industry.
Related variables include net mortality, strategic alliances to meet the
challenges introduced by information technology, and restructuring or merger
and acquisition activity directed toward the control of critical information
resources, such as vertical information integration. Although there is a lot of
anecdotal evidence on these impacts, there is little rigorous research in the area.

Industrial conduct

Research on efficiency-related impacts at this level would focus on comparative
efficiency, seeking to establish that technology adopters show efficiency gains
relative to their competitors, and on interorganizational synergies, attempting to
establish efficiency gains at the boundaries between organizations. Despite the
similarity to the study of efficiency impacts at the organizational level, the need

17

for a comparative emphasis, the consideration of inter-firm transaction costs, and
the extention of the unit of analysis beyond a single organization, complicates the
problem of the design of appropriate studies.

Although some of the variables underlying market power or sustainability have
been the subject of econometric studies (e.g., market share), and thus have
existing operationalizations, the majority requires the development of both
theoretical models for the impact of information technology and adequate
measures. Similarly, although the potential of the technology to promote
strategic alliances or mergers and acquisitions, and to afTect the ability of a firm
to adopt specific strategies is well recognized, we are not familiar with any efforts
to theoretically predict or empirically establish the nature of these impacts.

MEASURING THE UKi'ENUENT VARIABLES

Measurement and operationalization of theoretical variables is a self-contained
problem, as the development of appropriate instruments is closely related to the
theoretical and construct validity of the models tested. A discussion of these
issues extends beyond the scope of this paper, but Bagozzi (1979) provides an
excellent discussion for the interested reader. In the rest of this section we review
representative measures for the dependent variables identified earlier, and which
may be applicable to the study of organizational impacts of information techno-
logy.

Firm-level variables

A number of structural variables at the firm level have been used in empirical
studies. Organization theory provides classification schemes for organizational
forms, such as functional vs. product hierarchies vs. matrix forms. Vertical
integration has been measured as the ratio of value added to sales, although this
measure is biased towards raw materials producers. Several measures of
centralization have been used based on the locus of decision making, such as
profit vs. cost centers, budgetary authority, and number of reporting levels. The
number of job titles has been used as a measure of differentiation. Diversification

18

measures in the strategic literature are primarily based on the ratio of dominant
product line sales to total sales.

Process efficiency has been measured by using frontier analysis or DEA (data
envelopment analysis) on production cost data, which at closer scrutiny often
include coordination costs as well. Stabell ( 1983) and Chismar and Kriebel ( 1985)
conducted representative studies in the MIS literature. Other process variables
do not have existing operationalizations, and thus empirical researchers will
have to start from the reference discipline of organization theory and the MIS
literature surveyed in the previous sections.

There are several performance measures in the literature, ranging from financial
returns on investment and assets, to the value added as refiected by a firm's
market valuation, to direct cost/benefit assessments of the value of information
systems. These measures have the advantage of high validity, but as discussed
earlier, confounding factors and extraneous variance make the task of estab-
lishing an impact due to information technology extremely difficult.

Industry-level variables

Internal efficiency at the industrial structure level can be measured similar to
firm-level efficiency. Comparative measures of the efficiency of multi-product
and different size firms have been used to estimate the economies of scale and sco-
pe for individual industries, while inter-process inventories, delivery times and
purchasing costs have been used as a measure of interorganizational synergies.

The study of the structure and efficiency of markets has been an area out of the
mainstream of economics and organization theory. Market power and the degree
of collusion have been studied extensively, however, with empirical proxies such
as concentration ratios, existence of price wars, price stability, ability to price
over marginal costs, and competitiveness in advertising. Yet recent advances in
industrial economic theory have cast doubt on the ability of these measures to
capture tacit collusion.

Industrial organization theory is actively developing new theoretical models for
the barriers of entry, substitution and the role of reputation, which may lead to

19

related empirical measures. The rate of innovation has been typically measured
by patent counts and the number of new product introductions, measures often
unreliable and inappropriate for inter-industry comparisons. Macroeconomic
data is available on firm mortality in different industries and on mergers and
acquisitions, while industry performance measures are similar to the ones used
for individual firms. Once again, while the measures available are quite reliable,
they are not sensitive to the impact of information technology.

For variables specific to industrial conduct, market power has been measured in
terms of market share, excess profits, and ability to price over marginal costs. A
number of empirical studies have looked into asset specificity and customer
switching costs in specific settings to test the predictions of transaction cost
theory, but we are not aware of any generally applicable measures. Measurement
of innovation and vertical integration were discussed earlier, and we are not
familiar with applicable measures for the remaining variables in Table 3.
Interested researchers will have to develop appropriate operationalizations
starting from the theoretical literature.

CONCLUDING REMARKS

Not enough of the existing literature on the impacts of information technology at
the organizational and industrial levels consists of rigorous theory building or
empirical studies. Research results are fragmented and conflicting (Attewell and
Rule 1984). Most important, we cannot claim substantial progress in our
understanding of the organizational and industrial impacts of information
technology and in our ability to establish their existance and quantify their
nature, compared to early work in the organizational and strategic disciplines
(e.g., Zannetos 1968, 1970). We believe that an agreement on the appropriate
dependent variables will make a contribution to research in this area. The
underlying theoretical disciplines of organization theory, microeconomics and
industrial economics can provide us with sets of organizational variables, among
which the ones relevant to information systems research can be identified. We
used the framework shown in Figure 1 to organize the major firm and industry-
level approaches in these reference disciplines and to propose appropriate
variables.

20

By choosing appropriate explanatory theories from the underlying disciplines,
information systems research can focus on the process-related impacts of the
technology, relying on existing theory to link these impacts to organizational
performance (see Figure 2). There are a few salient characteristics of information
technology that can be used as the basis to study its process impacts and to
identify these variables that the technology is most likely to afTect. It has been
proposed that bounded rationality can provide such a basis from a rationalist
perspective of individual and organizational action, while organizational
flexibility is most relevant from an evolutionary perspective.

Weick (1984) pointed out the importance of proper methodology in studying the
impact of information technology on organizations, but he also cautioned us about
the dangers involved in trying to distinguish between independent and dependent
variables. In particular, as Markus and Robey (in press) explain in detail, there
are many bi-directional interactions in organizational systems. Causal models
will be inadequate if they do not include the appropriate feedbacks, and cross-
sectional data will be unable to show the interactions among the variables; hence
longitudinal approaches will be necessary.

As the level of analysis becomes more general (from the organizational unit to the
industry, to the economy and society as a whole), the links between independent
variables related to information technology and the dependent variables of
interest become more indirect, introducing a large number of confounding
variables and extraneous variance that reduce our ability to observe the impact of
information technology. In most industries the participant firms employ similar
types of information technology, thus reducing the variance of our independent
variables. Furthermore, several empirical studies in the past were based on data
from the impact of older technologies, such as accounting or payroll systems,
whose impact on organizations and industries is likely to be quite different from
modern user-friendly applications emphasizing analytical tools, communications
and specialized areas of strategic importance.

Developing valid operationalizations and instruments for any of the variables
proposed in this paper is a formidable task by itself. Providing empirical research
guidelines for the entire set we identified goes beyond the ambitions of this paper;
our goal was to point in the right direction. We conclude with a list that may

21

prove useful in addressing the complexities of empirical research into the
organizational and industrial impacts of information technology:

• We need to build a cumulative research tradition on the impacts of
information technology on different organizational variables and at
different organizational levels.

• We need more sophisticated theories (models), and methodologies (instru-
ments) to overcome the problem of confounding variables; industrial
economics can suggest apprpriate models, but to this date organization
theory does not offer sufficient assistance, as it is not able to provide
adequate opera tionalizations for several of its constructs.

• We must identify the variables on which the technology is likely to have
more direct impacts, e.g., variables that are closely affected by bounded
rationality at the individual and organizational levels.

• We should use methodologies that allow us to control or compensate for
the contextual variables when it is feasible; thus we can hope to better
isolate the impacts of information technology.

• We should study technology innovators and collect longitidunal data, so
that we can achieve more variance in our independent variables.

• Finally, we should study organizations in which information technology is
closely related to their core technology, and thus is more likely to result in
impacts easier to observe, as is the case in several service industries.

22

References

Argyris, C. "Organizational Learning and Management Information Systems."
Data Base, Winter-Spring 1982.

Astley, G. A. and Van de Ven, A. H. "Central Perspectives and Debates in
Organization Theory." Administrative Science Quarterly, Volume 28, June 1983,
pp. 245-273

Attewell, P. and Rule, J. "Computing and Organizations: What We Know and
What We Don't Know." Communications of the ACM, Volume 27, No 12,
December 1984.

Bagozzi, R. P. The role of measurement in theory construction and hypothesis
testing: Toward a holistic model. In Conceptual and Theoretical Developments in
Marketing, American Marketing Association, Chicago, II., 1979.

Bakos, J. Y. "Toward a More Precise Concept of Information Technology."
Proceedings of the Sixth International Conference on Information Systems, L.
Gallegos (ed.), Indianapolis, Indiana, December 16-18, 1985, pp. 17-24.

Bakos. J. Y. and Treacy, M. E. "Information Technology and Corporate Strategy:
A Research Perspective." MIS Quarterly, Volume 10, Number 2, June 1986.

Beath, C. M. and Ives, B. "Competitive Information Systems in Support of
Pricing." MIS Quarterly, Volume 10, Number 1, March 1986, pp. 85-93.

Blau,P. M. Exchange and Power in Social Life. Wiley, New York, 1964.

Bostrom, R. P and Heinen, J. S. "MIS Problems and Failures: A Socio-Technical
Perspective." MIS Quarterly, September 1977, pp. 17-32.

Carter, N. "Computerization as a Predominate Technology: Its Influence on the
Structure of Newspaper Organizations." Academy of Management Journal,
Volume 27, Number 2, June 1984, pp. 247-270.

Cash, J. I. and Konsynski. B. R. "IS Redraws Competitive Boundaries." Harvard
Business Review, Volume 63, Number 2, March-April 1985, pp. 134-142.

Chandler, A. D. Jr,. Strategy and Structure: Chapters in the history of the
American industrial enterprise. MIT Press, Cambridge, Ma., 1962.

Chismar, W. G. and Kriebel, C. H. "A Method for Assessing the Economic Impact
of Information Systems Technology on Organizations." Proceedings of the Sixth
International Conference on Information Systems, L. Gallegos (ed.), Indianapolis,
Indiana, December 16-18, 1985, pp. 45-56.

Ciborra, C. U. "Reframing the Role of Computers in Organizations: The
Transaction Costs Approach." Proceedings of the Sixth International Conference
on Information Systems, L. Gallegos (ed.), Indianapolis, Indiana, December 16-18,
1985, pp. 57-69.

Cohen, M. D., March. J. G. and Olsen, J. P. "A Garbage Can Model of Organiza-
tional Choice." Administrative Science Quarterly. Volume 17, 1972, pp 1-25.

Crawford, A. B. Jr. "Corporate Electronic Mnil: A Communication-Intensive
Application of Information Technology." MIS Quarterly, Volume 6, Number 3,
September 1982, pp. 1-13

23

Cron, W. J. and Sobol, M. G. 'The Relationship Between Computerization and
Performance." Information and Management, Volume 6, 1983, pp. 171-181.

Crowston, K.. and Treacy, M. E.. "Assessing the Impact of Information
Technology on Enterprise Level Performance." Proceedings of the Seventh
International Conference on Information Systems, San Diego, California,
December 14-17, 1986.

Cyert, R. and March, J. A Behavioral Theory of the Firm. Prentice-Hall,
Englewood Cliffs, New Jersey, 1963.

Feldman, M. S. and March, J. G. "Information in Organizations as Signal and
Symbol." Administrative Science Quarterly, Volume 26, 1981, pp. 171-186.

Freeman, J. "Organizational Life Cycles and Natural Selection Processes." In B.
M. Staw and L. L. Cummings (eds.). Research In Organizational Behavior, JAJ
Press, Greenwich, CT, Volume 4, 1982, pp. 1-32.

Galbraith, J. Organization Design. Addison-Wesley, Reading, Ma., 1973.

Galbraith, J. De'^igning Complex Organizations. Addison-Wesley, Reading,
Massachusetts, 1977.

Ives, B. and Learmonth, G. P. "The Information System as a Competitive
Weapon." Communications of the ACM, Volume 27, Number 12, December 1984,
pp. 1193-1201.

Keen, P. G. W. "MIS Research: Reference Disciplines and a Cumulative Tradi-
tion." Proceedings of the First International Conference on Information Systems,
C. Ross (Ed.), Philadelphia, Pennsylvania, December 8-10, 1980, pp. 9-18.

Kling, R. "Social Analyses of Computing: Theoretical Perspectives in Recent
Empirical Research." Computing Surveys, Volume 12, Number 1, March 1980.

Kriebel, C. H. and Moore. J. H. "Economics and Management Information
Systems." Data Base. Fall 1982, pp. 30-40.

Kriebel, C. H. and Raviv, A. "An Economics Approach to Modelling the
Productivity of Computer Systems." Management Science, 26, 3, March 1982.

Kuttner, R. "The Poverty of Economics." r/iei4//aahcA/on//t/y, February 1985.

Malone, T. W. "Organizational Structure and Information Technology: Elements
of a Forma! Theory." Working Paper 130, Center for Information Systems
Research, Massachusetts Institute of Technology, August 1985.

Malone, T. W. and Smith, S. A.. "Tradeoffs in Designing Organizations:
Implications for New Forms of Human Organizations and Computer Systems."
Working Paper 112, Center for Information Systems Research, Massachusetts
Institute of Technology, March 1984.

March, J. G. "Bounded Rationality, Ambiguity, and the Engineering of Choice."
Bell Journal of Economics, Volume 9, 1978, pp. 587-608.

March, J. G. and Olsen, J. P. Ambiguity and Choice in Organizations. Bergen,
Norway, 1976.

Markus, M. L. "Power. Politics and MIS Implementation." Communications of
the ACM, Volume 26, 1983, pp. 430-444.

24

Markus, M. L. and Robey, D. F. "Information Technology and Organizational
Change: Causal Structure in Theory and Research." Management Science, in
press.

Mason, R. O. and Mitroff, 1. 1. "A Program for Research on Management
Information Systems." Management Science 19,5 (January 1973), pp. 475-487.

Parsons, G. L. "Information Technology: A New Competitive Weapon." Sloan
Management Review, Volume 25, Number 1, Fall 1983, pp. 3-14.

PfefTer, J. 'The Micropolitics of Organizations." In M. W. Meyer and Associates
(eds.), Environments and Organizations. Jossey-Bass, San Francisco 1978, pp. 29-
50.

Pfeffer, J. Power in Organizations. Pitman, Marshfield, MA, 1981.

Pfeffer, J. Organizations and Organization Theory. Pitman, Marshfield, MA,
1982.

Porter, M. Competitive Strategy. Free Press, New York, New York 1980.

Robey, D. F. "Computer Information Systems and Organization Structure."
Communications of the ACM , Volume 24, Number 10, October 1981, pp. 679-687.

Salancik, G. R. and Pfeffer, J. "Who Gets Power - And How They Hold On To It: A
Strategic Contingency Model of Power." Organizational Dynamics, Volume 5,
1977, pp. 3-21.

Stabell, C. B. and Foresund. F. Productivity effects of computers in
administration: an exploratory empirical investigation. Report to the Economic
Commision on Europe, United Nations, Rome, 1983.

Thompson, J. D. Organizations in Action. McGraw-Hill, New York, 1967.

Trist, E. L. "The Sociotechnical Perspective." In A. H. Van de Ven and W. F.
Joyce (eds.). Perspectives on Organization Design and Behavior, Wiley and Sons,

New York 1981, pp. 76-87.

Tushman, M. L. and Nadler, D. A. "Information Processing as an Integrating
Concept in Organizational Design." Academy of Management Review, July 1978.

Vroom, V. H. Work and Motivation. John Wiley and Sons, New York 1964.

Weick, K. E. "Cognitive Processes in Organizations." In B. M. Staw (ed.),
Research in Organizational Behavior, Volume 1, JAI Press, Greenwich, CT, 1979,
pp. 41-74.

Weick, K. E. "Theoretical Assumptions and Research Methodology Selection." In
F. W. McFarlan (ed.). Harvard 75th Anniversary MIS Research Colloquium, 1984,
pp. 111-132.

Williamson, O. Markets and Hierarchies. Free Press, New York, New York,
1975.

Zannetos, Z. S. "New Directions for Management Information Systems."
Technology Review, Volume 71, Number 1, October-November 1968.

Zannetos, Z. S. "Computerized Management Information Systems and
Organizational Structures." Proceedings of the International Symposium on MIS,
University of Cologne, July 1970.

25

o

38

LIBRARIES

ill

3 ^060 0056S101 ?

D a t e D u e f !f-?D

OCT, 221992

Lib-26-67