HD28
.M414
WORKING PAPER
ALFRED P. SLOAN SCHOOL OF MANAGEMENT
The Politics of Technological Change:
An Empirical Study
Robert J.Thomas
July 1988
WP# 2035-88
MASSACHUSETTS
INSTITUTE OF TECHNOLOGY
50 MEMORIAL DRIVE
CAMBRIDGE, MASSACHUSETTS 02139
The Politics of Technological Change:
An Empirical Study
Robert J. Thomas
July 1988 WP# 2035-88
Contents
Introduction 1
Decision-Making around New Technology 5
Research Sites and Methods 7
The Flexible Machining System 10
A First Look at Outcomes 1 1
From Concept to implementation 13
The Framing Contest 15
Aligning Objectives 18
Framing Outward, Upward and Downward 20
Summary 24
Shop Programmable Machine Tools 25
A First Look at Outcomes 26
From Concept to Implementation 28
Aligning Objectives 30
Framing Downward, Not Outward 31
Summary 34
Implications for the Two Perspectives 35
Broader Implications 37
Lateral Dimension 37
Hierarchical Dimension 40
Footnotes 42
References 44
Introduction
Developments in manufacturing and information technologies over the past decade have
given rise to a substantial body of research and debate about the implications of technological
change for employment, skill levels, and economic competitiveness. On one side,
"postindustrial" and socio-technical theorists suggest that numerous factors - including
technological developments themselves - are presenting business organizations with the
opportunity (or, in the minds of some, the necessity) to fundamentally recast themselves in the
direction of greater flexibility and responsiveness to market conditions. These attributes, they
suggest, can only be achieved through a sea change in organizational approaches to the design and
deployment of technology and human resources (cf., Piore and Sabel 1984; Walton 1987; and
Hirschhorn 1984). It is argued, for example, that the potential benefits of new "informating"
technologies (Zuboff 1982) can only be realized through the grant of greater autonomy and
responsibility to employees formerly handcuffed by Taylorist definitions of automation; and,
according to Hirschhorn (1984) and others [most notably Davis and Taylor (1976)], the
greater complexity and sensitivity of new manufacturing technologies actually demand greater
- not lesser - involvement, commitment, and skill from production workers. In short,
technological and organizational "possibilities" are increasing in number. What remains at
issue is whether managers (especially American managers) will grasp those possibilities.
On the other side, analysts working from marxian labor process theory suggest that
neither technological developments nor hopeful scenarios are likely to shake managers from a
philosophy and a practice aimed at cost reduction, deskilling, and enhancement in control over
the work process. They point to the shopfloor use of new technology and question whether
anything has changed - even with the appearance of so many possibilities. In particular,
Braverman's (1974) depiction of the "control imperative" argues that managers use technology
to deskill the labor force and thus render it more subject to external control. In the wake of his
work, a host of case studies have been compiled which, while not necessarily refuting the
Politics of Technological Change
postindustrialists' reports of new developments, certainly cast shadows on the latters' claims of
the urgency or necessity of a philosophical and practical shift in management. Noble (1984),
for example, documents the winnowing of possibilities that took place with the development of
numerical control equipment and argues quite adamantly that options which might have allowed
machinists greater influence in the conduct of their work - a frequently cited possibility
attributed to new technology (cf., Hirschhorn 1984) - were discarded in the interest of
greater managerial control over the labor process. Shaiken (1986) and Howard (1987) reach
similar conclusions in their overviews of technological change in manufacturing industries.
To date, the evidence about job loss and creation, skiiling and deskilling, and
competitiveness and stagnation has not yielded anything resembling a clear pattern of outcomes
(cf.. Hunt and Hunt 1 985; Francis 1 986; Jones 1 982; Kelley 1 986; Spenner 1 987; Piore and
Sabel 1984). Part of the confusion in the debate over the consequences of new technology, as
well as the variability in choices organizations have, resides in the complex definitional and
methodological issues the topic surfaces: how jobs are counted, how skill changes are measured,
and how competitiveness - organizational or industrial - is operationalized.
But, equally imporiant, the debate over what is happening - much less the debate over
what might happen - has also been hampered by a relative lack of attention to the actual
processes by which organizations select and implement new production technologies. Failure to
thoroughly investigate the process and the context of organizational decision-making around new
technology has led to a situation in which a great deal is assumed (by theorists on both sides)
about: (1) managerial rationality or irrationality in problem-selection and problem-solving;
(2) the relative influence of economic and political/ideological criteria in decision processes;
and (3) the intentions or objectives which decision-makers seek to fulfill through technological
change.
In this paper, the results of two comparative case studies of technological change in a
major manufacturing company - a flexible machining cell and shop programmable machine
Politics of Technological Change
tools - argue that the process whereby organizations select and implement new production
technologies is an inherently political one. But, by contrast to prior characterizations (e.g.,
Pettigrew 1973, Dean 1987), I will argue that the politics of decision-making ought not be
seen as limited to the process of jockeying, lobbying, coalition-building, etc., associated with
securing approval for (or denying) a change proposal. To be sure, the case studies will show
that the framing of proposals and the justification of change are rife with partisan activity and
frequently contradict the tenets of rational behavior put forth by neo-classical economics (cf.,
Simon 1 955; March 1 978). However, there is another level of politics below the surface
which is intimately tied to the structure of power and authority in organizations. This level of
politics may prove no more rational (at least in the dictionary sense); but, I will argue, it does
provide a plausible alternative to both the indeterminacy of simple models of "pluralism" in
organizational politics and the over-determinism which tends to characterize most renderings of
labor process theory.
This structural dimension to the politics of technological change is most clearly revealed
in two stages. The first stage involves an analysis of the objectives or interests which underlie
proposals for (and opposition to) change. Three categories of objectives will be put forth as
critical: (1) personal obiectives or the interests which individual proponents and opponents
seek to promote or protect in the change process; (2) positional obiectives or interests
associated with functional distinctions within organizations (e.g., between production and sales
units; top, middle and lower management; or management and labor); and (3) o'^i^anizational
obiectives or interests emblematic of the enterprise as a whole and its relationship to the
external environment, including its competition.
Technological change options, I will argue, are filtered through a three dimensional screen
created by personal, positional and organizational objectives. Independent of whether the
proposed change is generated internal to the organization or drawn from the outside (e.g., from a
vendor), it will not pass through the screen (i.e., be implemented) unless there is at least
Politics of Technological Change
minimal alignment among those three objectives. Thus, we can expect that a great deal of effort
associated with technological change will involve proponents' attempts to create the appearance,
if not the reality, of alignment. Most commonly, proposals will be framed in terms of their
contributions to organizational objectives precisely because those objectives will constitute the
language of discourse within the organization itself. Personal and positional objectives will
either be presented as subordinate to overarching organizational objectives or as dwarfed by
comparison to the overwhelming significance of the organizational objectives. This constitutes
the core of what Dean (1987) has referred to as the "justification" process.
The second stage, however, involves an analysis of the models which underlie the
objectives and the relationships which influence the choice of models. Since the justification
process is only activated after problems are "found" (Pounds 1969), it is essential to
illuminate the models of present and desired states which guide what Lyies and Mitroff (1980)
refer to as the process of "problem formulation" for different categories of organizational
actors. In other words, I will suggest that the analysis of proposals for change must be
accompanied by an analysis of the more general models which guide or direct the identification of
the problem itself. However, problem formulation does not occur in a vacuum. It involves
reflection on the activities and behaviors of other organizational actors - be they subordinates,
peers or superiors. More importantly, aside from the purely technical aspects of production
(e.g., the characteristics of the raw materials undergoing transformation), alterations in
production technique invariably involve changes in the behaviors of human beings. The key
"problem" which anchors this analysis of models is what I will refer to as the "human problem
of transformation", i.e., the transformation of the "potential to work" which organizations
acquire when they hire employees and the actual output (in products or services) which
employees create.
Thus, 1 will argue, relationships among categories of organizational actors, particularly
power relationships, and ideoio^ical positions , particularly differences in perspective on what
Politics of Technological Change
motivates human beings, will directly influence the identification of problems and the choice of
solutions. By contrast to the postindustrial perspective, I will suggest that the proliferation of
"technological possibilities" alone is unlikely to alter managers' models of the human problem of
transformation - that will require a substantial shift in both ideology and practice. But, by
contrast to the labor process perspective, I will suggest that the human problem of
transformation is not limited to production work or production workers - in fact, it is often
the basis to top management's model of the rest of the organization , including line management
and staff.
Following a brief discussion of the rationale for studying technological change at the level
of the firm, I will describe the research sites and the methods employed in the study.
Decision-Making around New Technologv
Research and theory-building of the sort described in this paper hinges on two core
assumptions: (1 ) Organizations selectively respond to change pressures emanating from within
and from without. Ecological (Hannan and Freeman 1974), institutional (Zucker 1977 ;
DiMaggio and Powell 1983) or systemic constraints (Edwards 1979; Gordon, Reich and Edwards
1982) may limit the array of responses available to a given firm at a given point in time; yet,
organizations can and do select responses and transform them into empirically observable
actions. (2) Characterization of the decision-making process as "organizational" does not imply
that decisions (especially those which result in concrete actions) require either the
participation or the consensus of all organizational actors. Indeed, the exclusion of some
categories of actors is an important a part of the politics of decision-making (cf., Pettigrew
1973; Pfeffer 1977). However, by contrast to cognitive or psychological approaches to the
study of decision-making (cf., Kleinmuntz 1 987), the organizational aspect is important in that
it forces us to take into account both the process and the context in which change decisions are
made and it suggests that the organizational decision process, especially in large-scale firms,
cannot be modeled in the same fashion as that of the individual decision-maker.
Politics of Technological Change
There are three specific advantages to the study of decision-making around new technology
at the level of the firm.
First, it allows for the possibility that firms not only are capable of recognizing
alternative technological possibilities but are able to select among those possibilities in line
with a conscious or semi-conscious philosophy of work organization. At the same time,
however, it allows for the possibility that decisions about technology are not strictly rational in
the neo-classical economic sense - that is, they are characterized by a bounded or limited
rationality (March 1978) of a form which is essentially political (as the labor process
theorists suggest) or philosophical (as the postindustrialists suggest). The nature of those
boundaries or limits ought to be visible in the arguments (and data) marshalled in support or
opposition to a proposed change. This firm-level approach to the study of technological change
builds off earlier work in the politics of organizational decision-making, especially that of
Pettigrew (1973) and the more recent study by Dean (1987). However, the approach I will
use differs in one important respect: rather than conceptualize the "decision to innovate" (Dean
1987) primarily as a contest among individual managers or groups of line and/or staff
employees, 1 will seek to uncover the "problems" - discrepancies between the managers' "^^^
models and organizational realities (Pounds 1969) - which give rise to proposals for
technological change.''
Second, a focus at the level of the firm throws into relief the process through which
technological possibilities are surfaced ;e.g., how information is collected, filtered, and bundled
into proposals for change). If, as postindustrial theory suggests, the array of possible
approaches to work organization is expanding as a result of new developments in technology,
then it will be of more than passing interest to see where and how knowledge of technique or
precedent enters organizations. Sy contrast, if we follow labor process theory, the information
collection and filtering process is likely to be severely constrained to include only those
processes which allow visible cost reduction or some enhancement in management's ability to
control the production process. Filtering, in either case, ought to involve some assessment of
the probabilities of success and failure of particular change proposals (Cyert, Simon and Trow
1956). Thus, explanation of what determines the perceived probabilities of success in a given
change becomes a key to understanding the decision-making process.
Third, this approach forces us to investigate the process of technological change from
concept to and through implementation. Although arbitrary starling and ending points are
unavoidable and may underestimate the influence of past (or concurrent) decisions and
post-implementation alterations (cf., Rice and Rogers 1979; Kazanjian and Drazin 1986), the
emphasis on decision-making as a continous process of selection, testing and reformulation
makes it possible to avoid one of the most serious shortcomings of much of the work done in the
labor process tradition: the tendency to infer from outcomes to process. This weakness has been
particularly evident in the many of the case studies inspired by Braverman's popular work: a
tendency to undertake a cursory comparison between work organization under a new technology
with the system which preceeded it and to claim not only that skills and/or control have been
altered in the direction predicted by the theory, but that that outcome was precisely what
management wanted (cf., Thompson 1982; Child 1985; Kelly 1985; Wood and Kelly 1982).
One irony of this approach has been the tendency to attribute to management far greater
rationality, consistency, and, indeed, omniscience, than most management scholars (e.g.,
Drucker 1988) would be willing to cede.
Research Sites and Methods
To more thoroughly examine organizational decision-making around new production
technology, I undertook research in a Fortune 100 aircraft, aerospace and electronics
manufacturing enterprise in the summer of 1987. The company employs nearly 100,000
people in the United States and Canada has contracts with three major North American unions.
The union which also agreed to participate represents nearly two-thirds of the company's
hourly employees in the U.S. ^
Politics of Technological Change j
Two recent technological changes were chosen to represent new but substantially different
technologies being applied in the company's major manufacturing facilities.^ They were also
chosen for their "generic" qualities, i.e., their similarity to technologies being developed or
applied in a wide range of manufacturing enterprises. The projects became the vehicle for
analyzing the decision-making process. Although the project unit focuses attention on the
specific technology divorced from a stream of efforts of which it might be only one example, this
approach allowed us to investigate a case from beginning to end (i.e., from concept to and through
implementation).
Conditions of access made it necessary to choose technological changes which were, for the
most part, already completed. Since post-factum analyses can fall prey to the reconstructive
biases implicit in participants' recollections (cf., Schwenk 1985; Witte 1972; and Tversky and
Kahneman 1974), the research was conducted in such a way as to separately (a) document the
chronology of events surrounding the change in question and (b) query participants about their
role and their objectives in the process. More specifically, data collection consisted of three
main activities:
(1) In-depth semi-structured interviews were conducted with participants in the
decision-making and development processes. Initial interviews were conducted with
division-level research and development (R&D) managers to determine who was involved in the
various phases of each project. Subsequent interviews included additional people who were
mentioned by those interviewed on recommendation of R&D management. Thus, it was possible
to expand beyond the central engineering personnel to include representatives of allied functions
(e.g., facilities maintenance, materiel and purchasing, industrial engineering, training, and
industrial relations). Later interviews were conducted with corporate officials, line managers,
supen/isors, workers and union representatives in the facilities into which the projects were
placed. The interviews focused on the history of each project and on the direct or indirect
contributions each respondent made in that process. The interviews were tape-recorded and
Politics of Technological Change
transcribed. Each of the respondents was assured of anonymity in the reporting process and
every effort has been made to insure the confidentiality of the interview transcripts and the
comments, ideas, and aiticisms they contain. The factual content, as well as further
exploration of key items, was aided enormously by the opportunity to reinterview selected
respondents. In all. a total of 54 people were formally interviewed in the research. An
additional ten individuals were consulted informally or for short periods around specific issues.
(2) All available documentation associated with these projects was collected. This
material included project proposals and comments, memoranda and letters circulated internally,
purchase specifications, bids from equipment vendors, and capital equipment requests. Every
effort was made to corroborate dates, events, and (where appropriate) disputes by means of
existing documentation. On occasion this took the form of reviews of engineer's journals and
notes, i.e., logbooks which many engineers kept to document their progress, the hours they
devoted to given projects, and their comments on meetings and conversations. These latter
materials were especially useful in annotating and occasionally clarifying official reports and
schedules.
(3) Field notes were kept on the time spent in each site observing the equipment at work
and questioning the operators, supervisors and technicians in attendance about the technology,
particularly in terms of how it departed from past practices and how it affected adjacent
processes. Records were also kept on tours, hallway conversations and phone calls. Taken
together, these notes provided a measure of continuity to the analysis-in-progress and were
especially useful in generating new questions and lines of investigation. In addition, briefings
have been held (and will continue to be held) for the sponsors of the project. These briefings
and the feedback received in them has proven an important opportunity for learning on the part
of the sponsors and the author.
The cases will now be presented individually.
Politics of Technological Change 1 q
The Flexible Machining System
The automation of machining processes has been an area of great interest to engineers,
managers and academics for well over two decades (cf., Bright 1958). However, the creation of
flexible machining systems (or cells, referring to a series of machines linked to one another) is
a relatively new development. Until recently, the complexities and high cost of sensor
technology have made it uneconomic to automate a machining system unless the volume of parts
being produced was high and the variability of the cuts was low. The combination of automated
transfer equipment, remote sensing devices and computer-controlled or numerically-controlled
machine tools made sense for production runs of ten thousand to a hundred thousand automotive
crankshafts or engine blocks but not for 10 landing gear struts or 6 internal supports for the
tail section of a helicopter. Thus, in 1980, a machining system "flexible" enough to
economically produce several dozen different parts each with production runs of less than 50
parts per month represented a major departure from this company's prior experience and,
arguably, put it at or near the leading edge of such developments in the industry.
A cursory description of the technological possibilities alone offers few clues as to why it
might be attractive or how it might, in fact, be operationalized. Yet, a return to the two major
perspectives presented earlier gives us some rough hypotheses about the potential attractions
and the likely outcomes of a move in the direction of FMS technology. The FMS principals, in
particular, have been pointed to by Piore and Sabel (1984) supportive of an emerging strategy
of "flexible specialization" in manufacturing firms, i.e., the capacity to use the same equipment
to produce small quantities of highly varied parts to meet internal and market demands. This
strategy constitutes an about face from the familiar model of mass production with its heavy
emphasis on long production runs and low part variability. But, as Davis and Taylor (1976)
and Kern and Schumann (1 984) point out, flexibility has implications far beyond the technical
dimension. In particular, effective use of the technology requires the enhancement of employee
skills, i.e., the ability to monitor and direct expensive and often sensitive equipment becomes an
Politics of Teclinological Change 1 1
essential ingredient of the strategy. Moreover, the creation of a more tightly-coupled
production system brings with it what Perrow (1984) refers to as more complex interactions,
i.e., the potential for errors, accidents and/or malfunctions to occur as back-up systems and
adjacent (but indirectly linked) processes interact in unpredictable ways. The combination of
expense, sensitivity, and complex interactions suggest that employees should not only be skilled
but committed and responsible participants in the flexible machining system.
From the labor process perspective, automation - flexible or otherwise - has tended to
be seen as a linear extension of management's effort to reduce costs and expand control over the