Population Level
Learning and Industry Change
Edited
by Anne S. Miner and Philip C. Anderson
Advances in Strategic Management,
Volume 16
INDUSTRY
AND POPULATION LEVEL LEARNING: ORGANIZATIONAL, INTERORGANIZATIONAL AND
COLLECTIVE LEARNING PROCESSES
Anne S. Miner and
Philip C. Anderson
OVERVIEW
AND IMPLICATIONS
One important
school of strategic management theory has long adopted an implicit learning
perspective in describing how strategies are formed (e.g., Burgelman,
1988; Mintzberg & McHugh, 1985; Noda & Bower,
1996; Quinn, 1980). These authors have focused on the ways organizational
strategies emerge as individuals and groups come to learn about their own
situation and to converge on patterns of behavior that appear to work. Mintzberg, Ahlstrand and Lampel
(1998, p. 208-9) argue that the role of leadership is not to preconceive
deliberate strategies but to manage a process of strategic learning in which
novel strategies can emerge, noting that strategies often first appear as
patterns from the past and only later become plans for the future.
Even strategy
perspectives that emphasize formal analysis, planning and formal strategic
choice increasingly emphasize learning and knowledge deployment strategies.
Early research on the resource-based view of the firm (e.g., Wernerfelt, 1984) directed attention to knowledge-based
resources that are imperfectly tradable in factor markets. Those focusing on
the strategic management of technology argued that firm prosperity depends on
the proper assessment and deployment of knowledge assets, including both
information and competencies (Teece, 1986; Winter,
1987). These concerns naturally stimulated research on how such resources are
created (e.g., Leonard-Barton, 1995; Nonaka &
Takeuchi, 1995), and empirical studies of knowledge transfer between firms (Darr, Argote & Epple, 1995; Hamel, 1991; Ingram, 1997; Ingram & Baum,
1997), and in broader networks (Powell, Koput, &
Smith-Doerr, 1996). Indeed, the study of dynamic
capabilities lies at the heart of cutting-edge strategy research today (Iansiti & Clark, 1994; Teece,
Pisano & Shuen, 1997), focusing on organizational
systems for creating and harvesting knowledge that is difficult to imitate and
offers sustainable advantage. Links between learning and competition also inform
important work on the dynamics of international competition and industry
evolution as well as research on the co-evolution of technologies and
industries (Kogut, 1993; Teece,
1986).
At the same time,
learning research has also moved center stage in organization theory.
Dissertations and major academic journals increasingly highlight organizational
learning (e.g., Argote & Epple,
1990; Lant & Mezias,
1992; Levitt & March, 1988; March, 1991; Miner & Mezias,
1996; Walsh & Ungson, 1991). The growing body of
related empirical research builds on two identifiable traditions. The first
tradition emphasizes learning through the selective retention of standard
operating procedures or routines. In the traditional behavioral theory of the
firm, search for new routines is triggered by periodic performance gaps (Cyert & March, 1963, 1992; Lant,
1992), while other models emphasize less purposeful search and varied selection
processes (Burgelman, 1983; Miner, 1990, 1994).
Consistent with very broad models of adaptive systems (Holland, 1996; Huberman, 1989) the key learning outcome in this class of
models is change in behaviors of the learning system.
Another tradition
envisions learning as a cognitive process, involving inference, computation,
and the development of understanding (Glynn, Lant
& Milliken, 1994; Weick, 1979). Here, the key
learning outcome is change in knowledge structures or mental models. Recent
research has supplemented previous simulation (Levinthal
& March, 1981; Louanama & March, 1987) and
qualitative studies or conceptual studies (Cangelosi
& Dill, 1965) with systematic empirical research. Contemporary research
addresses learning subprocesses such as knowledge
creation, memory formation and access, search, forgetting and knowledge
transfer, and has provided evidence of distinct individual, group and
organizational level learning processes (Argote,
1999; Epple, Argote & Devadas, 1991; Levinthal &
March, 1993).
In this volume of Advances
in Strategic Management, we bring together a set of papers that link
strategy and learning models, with a special emphasis on how learning at
several levels of action produces change in collectivities of organizations. We
define a population level learning outcome as systematic change
in the nature and mix of routines, strategies or practices enacted in a
population of organizations, arising from experience (Miner & Haunschild, 1995). The population level learning framework
focuses on how repeated learning processes affect the distribution of routines
and competencies in an organizational population. These iterated learning processes
include recurrent (1) independent organizational, (2) interorganizational,
and (3) collective learning. The framework explores how iterated learning at
several levels of analysis will generate systematic population level change. We
believe many of the patterns discussed will have important implications for
organizational fields, communities and regional clusters, but emphasize
industry transformation in this volume and use population and industry
interchangeably.
Although each
paper contributes to its own distinct nexus of theory and practice, taken as a
whole they illuminate three central issues in strategic management:
In Table 1, we
indicate links between specific chapters and these issues, based on our
perceptions of the papers’ explicit and implicit implications.
Insight into these
three broad areas takes on increased urgency as organizations and industries or
organizational populations actively seek to use learning strategies and manage
knowledge in pursuit of survival and prosperity. These issues also resonate
with important theoretical questions regarding organization learning more
broadly and its links to strategic management. The framework may also have
special interest as organizational boundaries become more fluid, because it
focuses on practices and guiding recipes, rather than on numbers of organizations
in the population.
In this chapter,
we first clarify the population learning framework and definitions that inform
our approach, to lay the foundation for considering specific papers. Next, we
highlight selected ways specific papers provide insight into the three
questions above, noting implications for strategic management. To conclude, we
note selected additional themes raised by the papers, highlight areas we see as
ripe for further research, and sketch key contributions of the population level
learning framework to selected schools of strategic management and organization
theory.
LEARNING
OUTCOMES, PROCESSES, AND LEVELS
In this section,
we introduce our approach to learning outcomes, processes and levels, to
clarify the overall population level learning framework. We then build on these
ideas in the following section to argue that strategy development and
deployment unfold in the context of population level learning. After
summarizing several illustrative links between three key strategic issues
listed above and population level learning concepts and themes, we consider
each strategic issue’s links to the papers in more detail.
Learning as process and outcome. One reason
"learning" papers usually start by defining the concept is that many
plausible learning types and levels have been subsumed under this construct in
varied literatures. Despite – or perhaps because of – the burgeoning interest
in learning, the term itself has been somewhat contentious. Is cognition and
understanding essential for learning? Does knowledge creation represent
learning? At what level(s) of analysis does learning take place? Early writers
were frequently charged with committing the ecological fallacy, inappropriately
transporting concepts from one level of analysis to another and engaging in
reification of groups and organizations. Do organizations learn, or do only
individuals learn? Do organizations have memories?
We begin with a
basic, but inclusive definition of the process of learning: An agent
learns when experience systematically alters its behavior and/or its knowledge (Argote, 1999; Miner & Mezias,
1996). This definition characterizes both simple trial and error repetition of
apparently fruitful actions and complex intellectual discovery processes as
forms of learning. Consistent with this definition, we see a change in the
distribution of routines, practices or behaviors in the learning system as an
important learning outcome. This might, but does not necessarily, include
cognitive routines such as scripts, mental models or theories. Like March and
his colleagues, we define the learning outcome in terms of a change in
organizational routines or practices, not through outcomes such as greater
productivity (Levitt & March, 1988).
The link between a
context and behavioral routines evoked in that context may or may not, then, be
mediated by a knowledge structure. For example, an adaptive system that adjusts
future behavior through a simple feedback loop can be seen as ‘learning,’
although there is no notion of intellectual activity. On the other hand,
sometimes learning involves creating or updating a mental map, theories about
cause and effect, or taken-for-granted assumptions about constraints or
possibilities (see Walsh, 1995, for a review of managerial cognition and
collective memory). In this case, experience alters behavior by re-shaping the
connections in a knowledge structure, which can lead to subsequent change in
the distribution of behavioral routines, practices or strategies. The process
involves interaction with the world – or experience – with action informing
understanding and understanding potentially informing action (Daft & Huber,
1987).
Levels of learning. Viewing a learning outcome as change in the distribution of routines
and practices in a system puts a different light on concerns about levels of
analysis in learning. Rather than asking whether only individuals, or also
collections of individuals can learn, we point out that the distribution of routines, or patterned behavior can shift at many levels of
analysis. Clearly experience can systematically alter the future behavior
and/or knowledge structures of both individuals and collective actors.
In recent years
scholars have increasingly emphasized that learning at one level of analysis is
not merely the simple aggregation of learning at lower levels of analysis. Even
the individual mind may be thought of productively as a complex ecology of
interacting ideas (Bateson, 1972). Complex systems of interaction produce
emergent properties that cannot necessarily be predicted on understanding the
system’s components’ individual propensities (Cowan, 1994; Hutchins, 1991;
Simon, 1996). Learning at one level may be unrelated to, or even inconsistent
with effective learning at another level. For example, Argote
(1999) shows in detail how learning at the individual level can inhibit
learning at the level of social groups, and a simulation by Lounamaa
& March (1987) suggests simultaneous rapid learning by different parts of
organizations can reduce learning effectiveness. These considerations suggest
that collective learning is a worthy object of study in its own right. We
cannot always reduce a collectivity’s learning to the sum of learning processes
at the next lower level of analysis and action.
For this reason,
to understand learning as a driver and instrument of strategy, we not only need
to investigate lower levels of analysis such as learning by individual
strategists or top management teams learn (e.g., Huff, 1990; Reger, 1990). We must also examine the context provided at
higher levels of analysis. Organizational learning must be set in the context
of learning processes and outcomes at the next higher level of analysis,
involving multiple organizations that interact as they learn (March, 1991). The
emergent pattern of behavior that can become strategy at the organizational
level both influences and is influenced by the distribution of routines,
strategies and practices in the collectivity or population to which it belongs.
Expanding the definition of Baum (1996, p. 77) we define a population of
organizations as a set of organizations that share at least one major
semi-stable trait, activity or resource utilization pattern. An industry,
for example, would represent an organizational population; a regional
collection of organizations might also represent a local organizational
population.
Population level learning as outcome. Miner and Haunschild (1995, p. 115) originated the term population
level learning to denote the outcome of "systematic change in the
nature and mix of routines in a population of organizations as a result of
experience." Miner and Haunschild used the
term "routine" to refer to a broader class of activities than narrow,
repetitive sets of micro-actions in organizations. Their use of the term
included systematic bundles of activity such as practices, programs, search
rules, strategies, and other patterned behavioral or cognitive activity (Miner
& Haunschild, 1995).
These routines can
be enacted at many levels of activity; they might include organizational roles,
group practices, organizational strategies, or routines enacted at the level of
a population itself, such as an industry-level coordination ‘routine’ used in a
research consortium or trade association (Aldrich & Sasaki, 1995; Aldrich
Zimmer, Staber, & Beggs,
1994). Because it is based on experience, this learning outcome is distinct
from shifts in distribution of routines that are the result of differential births
and deaths of organizations. That is, the population level learning outcome is
not a compositional effect, in which some routines, practices and strategies
become more frequent because organizations which implement them become more
frequent. Rather, the change in distribution is itself the product of
experience of some unit or units in the population.
For example,
consider the U.S. radio industry in the 1960s and 1970s. AM
radio stations in the United States exhibited a distribution of programming
formats skewed heavily toward top-40-singles music. By the 1990s, most music
formats were offered by FM radio stations, and the distribution of AM formats
was skewed toward syndicated talk-radio programs. How did this shift in the
distribution of routines over time occur? In part, the transformation was
driven by differential births and deaths – founders of new AM stations tended
to adopt talk formats instead of music formats, and stations featuring music
formats tended to be sold to new owners who switched to talk formats. However,
this transformation is also in part a story of population-level learning. AM
radio stations learned from observing other stations that talk radio could
attract listeners and advertisers, so they changed routines. Additionally, supra-organizational
institutions emerged that made it easy for stations to institute talk radio
formats (in this case, widespread syndication and the emergence of recognized
talk radio personalities with nation-wide visibility) (Smith, Wright & Ostroff, 1998).
A population level
learning outcome, then, represents a sociological fact that characterizes a
collective entity, not its individual members. Individual or groups of
organizations enact routines; populations exhibit a distribution of routines.
To say that a population level learning outcome has occurred is to say that the
distribution of routines (practices and strategies) in a population has shifted
between two time periods, as a result of experience somewhere within the
population.
We emphasize three
learning processes that can produce this population level outcome, each at a
different conceptual level of analysis. First, parallel independent
organizational learning can produce new distributions of routines. As Carley shows (this volume), a set of organizations that are
initially similar can learn independently in the same context with the same
tasks, yet end up stratified in terms of practices and performance. Variety in
practices arises over time as the result of different individual organizational
experiences and learning trajectories.
Second, recurrent
interorganizational learning can produce new
distributions of routines. When an organization learns by observing
other organizations, by exchanging knowledge with others, or by generating
knowledge through joint interaction, this represents interorganizational
learning. This learning may be either strictly vicarious (when one organization
observes but does not interact with another) or interactive (when the learning
process arises from active contacts between two or more organizations).
Considerable research focuses on the impact of interorganizational
learning on the individual learning organization. Does an organization gain
from attempting to share knowledge in alliances (Aldrich & Sasaki, 1995;
Mowery, Oxley & Silverman, 1996; Powell et al, 1996)? Can a second mover
learn effectively from observing a first mover (Antonelli,
1990)? Yet it often stands silent on how repeated instances of interorganizational learning influence the shape of an
entire industry or organizational population.
Third, whole
industries and populations can observe other industries or populations, drawing
on their shared experience to enact or inform population-level routines. This
represents a collective population learning process at the population
or industry level. A collective population-level learning process occurs
when at least one learning step – such as knowledge discovery, retention,
retrieval from memory, or contemporary experience – is collective. For example,
a population/industry may observe its own shared experience and collectively
develop an inter-organizational coordination device (a population level
routine). It might also develop a formal collective entity after watching
another population/industry do so, build norms based on prior shared
experience, or evolve shared mental models of population/industry identity,
boundaries or performance standards. The United States semiconductor industry’s
observation of Japanese research consortia, and the consquent
development of its own research consortia (such as SEMATECH; see Corey, 1997)
represents a collective learning process. An emerging industry’s development of
a trade association group in response to the shared experience of an external
threat could also represent collective learning.
An important goal
of this volume is to explore the claim that learning processes at one level may
or may not mirror learning processes at another level. At the same time, it is
clear that the degree to which learning processes are collective represents a
continuous rather than a discrete variable. For example, if an industry
association (a collectivity) drew on data about industry level performance
(information from the memory of the whole population) to develop a new
coordination routine between industry members (new population level routine)
and codified it in an industry level rule (new form of population level
memory), an entire learning process would be collective at the industry level.
At the other extreme, population level learning might involve primarily one
collective learning step combined with other processes. For example,
organizations might pool their separate experiences with different
technological variants to decide on a shared industry standard. The original
knowledge acquisition at the firm level would have been organizational, but the
codification into industry-level norms to which all must conform would
represent a collective process.
While most papers
in this volume tend to emphasize one of the three levels of learning processes,
many also consider more than one process, and address partially collective
learning processes. Carley, for example, presents a
simulation study of parallel independent organizational level learning; papers
by Baum and Berta, Greve, Ginsberg, Larsen and Lomi, and Mezias and Eisner
primarily explicate processes of interorganizational
learning while Anderson, Lant and Phelps, Miner et
al., and Rura-Polley describe collective learning
processes. At the same time, these authors frequently show how learning
processes occur at more than one level, and most imply that different learning
processes influence population level transformation in non-obvious ways. The
power and complexity of learning’s impact on industries and organizational
populations can be overlooked when one focuses solely on strategy as
organization-level learning.
LINKS
BETWEEN POPULATION LEVEL LEARNING
AND
STRATEGY DEVELOPMENT AND DEPLOYMENT
Why should
students of strategic management care about population level learning outcomes
and the learning processes that produce them? Industry and competitive analysis
(e.g., Oster, 1999; Porter, 1980) represents a set of
frameworks that scholars and managers use to understand the context of an
organization’s strategic position. It helps us assess which industries are more
or less attractive environments, and how a firm’s strategy positions it with
respect to other enterprises. Much traditional industry analysis has drawn on
fairly static concepts of industry structure. The nature and mix of routines,
strategies and practices in a population that result from experience is the
context in which the strategy-making process takes place. By considering how
learning processes produce such changes, we begin to develop models of dynamic
transformation of the industry and population context for firm level
strategies. Theories about population level learning outcomes help us assess
why some organizational populations learn faster than others do, and how a firm
is positioned with respect to other enterprises whose experience influences its
own routines, strategies and practices.
In the following
three sections we consider specific implications of the papers in this volume
for each of the three major themes: (1) industry structure, (2) industry
evolution, and (3) industry attractiveness and robustness. To highlight some of
the common themes and implications of the chapters, Table 2 summarizes some of
the key insights developed in this volume that relate to each factor. Specific
rationales and nuances of related papers and arguments are considered in the
detailed sections in Table 2.
Industry
structure and learning
If firms are
repeatedly learning from their own experience and that of others, one might
expect populations and industries to become homogeneous in their routines,
strategies and practices. In the well-known evolutionary economics model of
Nelson and Winter (1982), for example, it is assumed
that each firm can copy any superior technology in the neighborhood of its current
techniques, and can imitate the practices of higher-performing firms, subject
to some error. If firms learn repeatedly from the same reality, or from each
other, how can they create and sustain competitive advantages over time in
knowledge-based competition?
Several papers in
this volume illuminate aspects of this central puzzle in strategic management.
They explicate how repeated learning at several levels can produce inter-firm
differences. Firms can become segregated into persistently different performance
categories characterized by different learning paces. Patches of collective
strategies can emerge that are isolated from each other, and populations can
break up into local learning neighborhoods. Importantly, heterogeneity does not
simply arise because of barriers to efficient or effective learning; it is
generated by the dynamics of learning processes themselves. Furthermore, in
some cases, learning produces population/industry level contours that shape
further organizational learning and population/industry transformation, so that
the trajectory of learning over time is heavily dependent on history. We
highlight in this section four specific factors that may influence industry
structure: recurrent independent learning, emergent
patches of collective strategies through cycles of competition and learning,
collective learning, and changing diversity in pools of routines.
Stratified industry/population through repeated independent learning. Carley (this volume)
describes results of a simulation study that models internal organizational
learning with individual level learning and organizational features that shape
individuals’ interactions. Sets of organizations begin with roughly similar
features, and face the same tasks. Although the organizations are all learning,
they do not converge to a single type of organization or set of organizational
features in the population as a whole. Repeated parallel independent
organizational learning (by organizations that do not learn by observing each
other) can lead to different strata, each containing a set of organizations
that is locked into a particular change strategy. Although there is oscillation
in specific rankings, the high-performing organizations in one period are also
typically the high performers in the next.
Carley observes that high performers in this stratified system do not
necessarily share fixed features with each other. Rather, their internal
learning processes and pace of internal change tend to be similar. Repeated
independent organizational learning creates a shift in the distribution of
organizational performance, traits, and behavioral learning routines in the
population. The latter appears to be most relevant to the performance outcomes
for individual organizations. The important general insight is that quasi-stable
industry patterns emerge not because some organizations are learning from their
experience while others are not, but because of detailed, local features of
learning processes within each organization. This in turn affects each organization’s
specific internal networks and procedures, which, in a path dependent process,
produces quasi-stable industry/population distributions of routines, strategies
and practices – along with persistent performance differences.
This simulation’s
results contradict the popular assumptions that repeated independent learning
by initially similar firms charged with the same task will necessarily produce
convergence to a shared set of practices and routines. It supports predictions
that firm level learning mechanisms tend to lock in, which in turn produces
population level patterns that are not easily altered. Firms can’t generate
superior performance simply by emphasizing the importance of learning, or by
imitating other population members whose results are superior. It implies
that learning firms need to attend to the fine-grained details of their
internal learning pace and processes if their learning is to provide
competitive advantage, and that they can expect high and low performance to
persist even all firms pursue a learning strategy.
Emergent
patches of collective strategies. In Ginsberg, Larsen & Lomi’s simulation
(this volume) firms compete in a rounds of a
prisoner’s dilemma within neighborhoods of interaction, then adopt the strategy
of the winner of the competition in their own neighborhood. This method of
determining their competitive strategy for the next round can be seen as interorganizational learning in the form of outcome
imitation (Haunschild & Miner, 1997). The
simulation explicates how repeated rounds of competition and interorganizational learning, under suitable conditions,
produce patches of organizations that follow a collective cooperative strategy:
the firms in each patch ‘lock-in’ to cooperation with each other. Their model
shows how local decision rules followed over time induce and sustain aggregate
regularities in the industry/population as a whole. Ginsberg et.
al. find that the initial proportion of cooperative
strategists in the population influences whether patches of interfirm
cooperation form. Increased scope of interaction produces larger islands of
cooperation in this context. They interpret this range of interaction in terms
of the cognitive limitations of participating organizations, but the range of
interaction could also be determined by such factors as shared sense of
identity among organizations or simple geographic proximity (Anderson, this
volume; Lant & Phelps, this volume).
Like Carley’s simulation, this paper shows how learning at a
lower level of analysis can produce regularities at the population/industry
level that then shape future interactions among
organizations. In these models, collective learning emerges from interaction.
Lower level learning produces stable contours at the population/industry level
that constrain the ongoing strategies and behavior of firms in the future, as
well as the learning strategies open to these firms. Again, these models imply
that firms pursuing a learning strategy may need to consider whether the
collective impact of firm learning may be creating broader industry/population
patterns that will constrain the focal firm in the future. For example, in
Ginsberg, Larsen & Lomi’s set-up, a firm that
prefers to maintain a cooperative strategy might intentionally look for
contexts that combine a high proportion of firms following a cooperative
strategy within one’s neighborhood, a larger scope of interaction and more
sensitivity to diffuse competition in the process.
Collective learning and population/industry heterogeneity. Anderson (this
volume) tracks actual changes in the distribution of two competing production
routines in the cement industry. He describes how the U.S. cement industry
moved from a period in which dry mixing dominated, through a period in which
wet mixing of raw materials was most common, and then returned to a
preponderance of dry mixing. Early in the period studied, that U.S. cement
industry developed a shared emphasis on quality as the key competitive
dimension. Customers had no reliable technical way to assess cement quality
directly, so they used social cues, reputation or high price as indicators of
quality.
After World War I,
a new wave of foreign competition led to a shared belief that thorough mixing
of ingredients produced better cement that achieved load-bearing strength more
quickly. Although this view was not grounded in direct scientific evidence, it
became a shared, taken-for-granted belief – derived through interorganizational
learning – that helped generate a shift to the use of water to gain more thorough
mixing at the cost of fuel efficiency and throughput. PCA-sponsored research
eventually generated ways to actually test quality (e.g., durability) of cement
directly, which in turn permitted technical standard levels to be adopted by
the industry. It also vastly improved dust control, eliminating a key advantage
of the wet-mixing process. Eventually, through further innovation, dry mixing
re-emerged as the dominant routine in the industry.
Both simple
imitative learning and a change in knowledge structures occurred. Indeed,
collectively sponsored research actually produced more accurate shared theories
about what produced more durable and reliable cement. More interestingly, the
original shared performance standards shaped collective and firm level learning.
But the results of this learning then transformed the industry’s performance
standard, which provided the new learning landscape for both firms and
collective learning. Which routine prevails as the distribution of routines
shifts as a result of experience-and which firms win and lose as their routines
fall into and out of favor--is a function of collective action, not economic
superiority alone. Firms can affect the outcome of
competition among clusters of routines by influencing how a population collectively
defines which dimensions of merit matter, what standards collective bodies set,
and whaT research trajectories such bodies pursue.
In another
approach emphasizing collective learning, Lant and
Phelps (this volume) argue that crucial learning occurs in the interactions
between the organizations (rather than only within the firms themselves). They
build on emerging sociocognitive theories of
strategic groups and on situated learning theory to consider this process of
collective population learning, drawing upon observations of the emerging media
industry of Silicon Alley in New York City. Lant and
Phelps go beyond the idea of shared mental models by conceptualizing
population/industry level learning as actually occurring within the ongoing
interaction of the firms over time. Lant
and Phelps’ approach implies that the development of cognitive groups does not
always result in the strategic myopia described in Scottish knitwear makers
(Porac, Thomas & Baden-Fuller, 1989). They suggest that variations in
perceived boundaries and strength of identity can generate heterogeneity within
cognitive strategic groups.
Changing pools of diversity of routines. Rura-Polley (this volume) tackles potential origins of
population/industry heterogeneity among practices by studying directly the
nature and degree of variation in proposed childcare practices in two distinct
organizational subpopulations: Catholic childcare institutions in Germany and
the United States over a 75-year period. She argues these two national contexts
produced very different patterns of variation among proposed practices, even
though the overarching institution – the Catholic Church – maintained its
underlying values and policies related to children’s institutions. Rura-Polley’s qualitative analyses indicate variation in
these candidate routines and strategies – at least in the German case –
reflects broader social trends. The degree and nature of variation in candidate
practices was clearly different in the two countries. Temporal patterns were not
the same, and the two countries differed in which topic areas showed most
variation. Rura-Polley’s work implies that the
candidate pool of routines/strategies available varies between subpopulations. This
suggests industry structure may arise in part from the degree of available
diversity of practices and strategists might consider the nature and pool of
available practices in choosing industries to enter.
Taken together,
these papers demonstrate ways in which all three levels of learning can, --separately and in combination – produce
population/industry heterogeneity, which in some cases may be sustained through
lock-in processes. One of the most important features of this insight, we
think, is the notion that while individual firms learn, they may be creating
unintended contours within their own industries that constrain their own fate
later on. Insight into the dynamics of such processes provides an
additional strategic factor for organizations whose fate will depend on the
survival and prosperity of their industries as a whole.
Industry
Evolution and Learning: Moderators of the Impact of Recurrent
Organizational,
Interorganizational and Collective Learning
Population-level
learning occurs when the distribution of routines – broadly defined – in
an industry changes as a result of experience. Industry evolution occurs when
the distribution of organizations themselves, or
organizational forms in an industry changes over time. For example, when
a change in the proportion of generalists versus specialists, M-form versus
U-form enterprises, or independent firms versus joint ventures persists over
time, the industry has evolved, with important strategic consequences. Since an
organization’s routines help govern its form and indeed some forms can be seen
as complex bundles of routines, it is clear that these two dynamic processes
are interrelated. Traditional industry-level competitive analysis ascribes
industry transformation to exogenous factors, such as technological or
regulatory change, and to the winnowing effects of competition. The articles in
this volume suggests that who organizations learn from and why they
adopt routines from each other is an endogenous factor that also
contributes to the transformation of a population over time.
The chapters
highlight features of learning processes that may shape industry evolution:
learning from near vs. learning from distal neighbors, within-group versus
across group learning; and modes of organizational learning.
Learning
from near vs. learning from distal neighbors Several papers explicate important ways
in which distance between organizations – whether in terms of space or other
dimensions – moderates the impact of learning processes on industry evolution.
For example, Greve studies organizations with multiple branches. The
evolution of the radio broadcasting industry in recent years has been marked by
a rise in the number of stations belonging to such multi-station,
multi-geography companies (Smith, Wright & Ostroff,
1998). Is this because branch systems facilitate the transfer of routines from
one region to another? In his empirical study of format changes in radio
stations in varied regions, Greve finds that greater
experience of a branch system’s units outside a given geographic market
causes lower performance of each branch in a focal market. Greve
also finds that the more experience a branch system has outside a given local
market, the greater the variation in performance in that market. The results
imply routines or format strategies transferred from distant neighborhoods
actually harm rather then help the imitating
stations. He argues that the difference in competitors between different
regions creates this effect. These findings imply that if firms repeatedly
imitate practices from distant neighborhoods, interorganizational
imitation may decrease the immediate performance of imitators. However, Greve points out bringing in routines from distant organizational
neighborhoods could help the local neighborhood as a whole, if other
organizations learn from the failed routines that local branches borrowed from
their sister stations.
Distal learning
also occurs among organizations that do not share common ownership. Baum and
Berta (this volume) study a group of firms competing in a behavioral
simulation. They find that organizations learn from other organizations when
their members interact with each other as individuals more frequently, and when
the imitating organizations have similar market shares. Interestingly, they
find that firms do not tend to imitate the most successful competitors in their
local populations. Rather, they look further afield, imitating the most
successful firms in other populations. Ginsberg, Larsen, and Lomi’s study suggests that the further away from their
neighborhoods firms search for successful organizations to imitate, the less
fragmented an industry becomes.
These studies
imply the boundaries of neighborhoods of interaction will powerfully influence
industry evolution, and that choice of neighborhoods of interaction as well as
imitation may represent an important strategic choice for individual
organizations.’
Within-group
versus across-group learning: Neighborhood boundaries and their permeability. The distance
between two organizations that may learn from each other is not simply a
continuous measure in social or physical space. Industries can fragment into
subgroups; firms that belong to different subgroups have to cross a boundary to
learn from one another, even if they are close demographically or
geographically. A change in the distribution of firms across such groups
constitutes industry evolution, if group membership is relatively distinct and
persistent. How are such changes influenced by learning processes?
Wade and Porac’s chapter in this volume builds a model suggesting
that the movement of personnel among firms in an industry is an important
conduit for interorganizational learning that
influences subsequent industry structures. Industry evolution in turn creates
opportunities that encourage or discourage certain types of managerial
migration among firms. In this way, interorganizational
learning (through the movement of people between organizations) and industry evolution
are linked in a mutually causal loop.
Early in an
industry’s history, these authors suggest, key managers from different firms
tend to migrate from the same industries, bringing with them a common set of
routines. For example, German experts greatly influenced the early practice of
both the American beer brewing industry and the American cement manufacturing
industry. Consequently, routines tend to be fairly homogenous across firms, and
collective structures that promote collective population-level learning emerge
fairly easily.
As an industry
evolves, Wade and Porac suggest, it tends to split into two subsectors, a core
and a periphery, differentiated by status, size, and/or performance.
Organizations belonging to the core tend to monitor and learn from each other.
Population-level learning takes place as routines diffuse through movement of
managers from core to peripheral organizations, and through experimentation
among peripheral firms, leading to the emergence of novel routines. When an
industry is jolted by a shock, personnel flows tend to reverse; core firms
begin to recruit from the periphery in order to import novel routines.
In considering
implications of their observations, Wade and Porac argue the degree of
permeability of an industry’s core may have important implications for the
industry/population’s long term adaptation. If core and periphery boundaries
are nearly impermeable, the core may fail to learn new routines, strategies and
practices, and over time fade. In contrast, if the core/periphery boundary is
permeable, personnel movement from the periphery may introduce routines and
practices that permit the to core to adapt through the
adoption of these new practices. Lant and Phelps make
a similar argument, pointing out that permeable boundaries between cognitive
strategic groups may permit new routines and strategies to move into
established groups, providing variance permits adaptive learning that can
enhance long term survival of the strategic group.
These theories
imply industry evolution will be influenced by permeability of boundaries
around subpopulations, and that important strategic action may include choices
about the strength and type of boundaries maintained.
Taken together,
these different moderating effects of neighborhoods on the population impact of
interorganizational imitation have important
strategic implications that flow from understanding who
organizations target for learning. Greve’s work
implies that learning from organizations too distant from oneself often reduces
performance, yet imitating nearby competitors can produce escalation in which
all firms ‘improve’ in absolute terms but not in relative position (the
so-called Red Queen effect in evolutionary theory) (Kauffman, 1993; March,
1991; Van Valen, 1973). This points
to the significance of carefully selecting interorganizational
learning targets. These localized neighborhood effects on repeated interorganizational learning also suggest that strategists
may benefit from considering not only their own targets for interorganizational
learning, but also the general pattern of interorganizational
knowledge flow in their own industry or strategic group. Superior selection
of individual interorganizational learning targets
could provide some firm level competitive advantage within an industry or
strategic group. At the same time, inadequate permeability of neighborhood
boundaries may reduce the viability of the whole industry/population.
Modes of interorganizational learning. Imitating other
organizations constitutes one form of interorganizational
learning: imitating another organization means using its experience as a basis
for one’s own actions. The process may or may not involve forming inferences or
causal theories about the imitated practice. On what basis do organizations
choose which routines to imitate? Three different modes of interorganizational
learning are addressed by various papers in this volume: (1) frequency based
learning (when organizations adopt the practice of a large number of other
organizations), (2) trait based learning (when organizations adopt the practice
favored by organizations with a distinct trait such as size or prestige), and
(3) outcome based learning (when organizations observe the apparent results of
a practice and adopt it or not based on those results) (Haunschild
& Miner, 1997; Miner & Raghavan, 1999).
Baum and Berta’s
behavioral simulation presents further evidence for the use of more than one
mode of interorganizational learning and imitation:
their groups appear to choose learning targets both on the basis of traits and
outcomes. Mezias and Eisner theorize that each
learning/imitation mode produces a different pattern of industry
transformation, due to differences in their relative speed and cost. They
contend that outcome-based learning leads to the most rapid evolution when
ambiguity is low, because successful new organization forms diffuse rapidly,
and firms quickly exit highly competitive niches. However, they believe the
costs of searching for a new form and adopting a new form are highest when
firms employ outcome-based learning. Combining intuitions about costs and
speed, Mezias and Eisner predict that both
frequency-based and trait-based learning will lead to long periods of stability
in the distribution of organizational forms, punctuated by short periods of
rapid, large-scale change. In contrast, outcome-based learning will result in
smaller, more frequent changes in an industry’s mix of organizational forms.
One implication of
work on modes of imitation and interorganizational
learning is that firms may profit from considering the broader pattern of interorganizational learning within which they function, so
they can anticipate where their industry/population may be moving, given that
repeated interorganizational imitation by multiple
organizations can produce unexpected effects. Firms may need to pay
attention to the interorganizational learning modes
of other organizations – not to understand what specific knowledge a competitor
may be gaining, but in order to gain some insight into how the collective
pattern of imitation might transform the industry’s practices over time.
Population
and industry attractiveness/robustness:
Learning’s
impact on population survival and prosperity
It is often
assumed if the individual organizations are all learning in sensible ways, the
aggregate result will be adaptive for the population or industry as a whole.
Yet the research in this volume shows this assumption to be false, and
explicates how different learning processes may interact to promote or detract
from population/industry survival and prosperity. We describe two important
factors the chapters imply will shape how learning influences survival and
prosperity of the whole population/industry: the degree and nature of tensions
between organizational and population level learning, and temporal learning
patterns.
Tensions between organizational and population level learning outcomes. Several papers
imply that lower-level learning may not only have unintended impact on
population/industry transformation, but may in some
cases conflict with fruitful population level learning outcomes. These studies
raise important questions about the impact that learning processes have on
organizational and population/level performance, including both survival and
prosperity.
Miner, Kim, Holzinger and Haunschild draw on
historical reports of organizational populations and industry transformations
to examine ways failure and near-failure by individual organizations may spur
different degrees of collective learning processes. The authors describe
learning processes in which failing organizations stimulate other organizations
to avoid specific routines/strategies/practices, to search or imitate nonfailing organizations or to reinforce their own current
practices. They also describe collective learning processes in which the
population/industry develops new collective routines such as research consortia
or federated political groups or spur compensatory actions.
Miner et al. argue
these processes produce shifts in the distribution of routines/strategies in a
population in several nonobvious ways. For example,
they observe how dominant firms may imitate challengers in the periphery of an
industry and fail in their own revival while legitimating the routines they imitate.
This does not imply that all learning from the failure of others is valuable.
However it suggests that the failure or near failure of some organizations in
many cases does spur fruitful learning by other organizations and the
population/industry as a whole. In fact the authors argue the
population/industry may learn more from failures or near failures than from
apparent successes, because of their greater visibility and potential for
developing more accurate causal models.
Similarly, Greve shows in his radio study how organizations that
imitate practices from distant regions reduce the level and increase the
variance in their own performance. He suggests this helps the
population/industry to which they belong by introducing new routines/strategies
and allowing others observe the impact of these new routines/strategies. He
argues such firms engage – in effect – in altruistic learning: their experience
detracts from their own performance but may enhance that of the
industry/population to which they belong.
Three other papers
point to the complementary situation in which fruitful firm-level learning
detracts from learning at the population/industry level. Lant
and Phelps, Wade and Porac, and Carley all describe
organizational learning processes that reduce variance within the firms
themselves. They point out that this useful firm-level learning may restrict
population/industry learning and adaptation by reducing the variation in
routines/strategies available in the future. These situations can be seen as the
reverse of the failure argument, which hinges on the idea that contrasts
between failure, near-failure and success provide the variance in both behavior
and outcome that represents the most valuable learning context.
At the heart of
all these studies is the idea that individual organizations face trade-offs
between potentially valuable organizational and population-level learning
processes. The general dilemma is familiar to firms in the area of knowledge
deployment in research consortia, where firms try to balance supporting
collective discovery of knowledge useful to the entire industry with
safeguarding their own knowledge and avoiding free-riding by others (Mowery,
Oxley & Silverman, 1996). The papers in this volume suggest the dilemma may
apply more broadly than in the obvious context of research and development.
Efficient solutions to this puzzle are not apparent. It is hard to devise
appropriate systems of side payments that reward failing firms for enhancing
the knowledge of others or deviant, low-performing firms for preserving
variance potentially useful to the whole population. It suggests that
collective consortia and industry associations could deliberately facilitate
diversity in the populations/industries they represent, rather than relentlessly
promoting ‘best practices’ in all members. At the same time, structural
features of the population’s competitive system can crucially influence the
best strategy for the firm’s learning, consistent with March’s (1991)
description of ecologies of learning.
Population/Industry performance impact of temporal learning patterns. March (Lounamaa & March, 1987; March, 1991) has illustrated
potential dangers of overly-rapid learning in simulations of organizations
learning from their own experience, especially in the presence of noise and
multiple learning entities. The papers here underscore potential tradeoffs
between fast and slow experiential learning at the population/industry level.
Baum and Berta’s
paper explores the important question of how speed of population level learning
affects industry performance and survival. Speed is an important issue in all
forms of learning (Argote, 1999). Baum and Berta
provide the first systematic empirical evidence supporting the prediction that
populations will face windows of opportunity in which learning is most fruitful
at the population/industry level (Miner & Haunschild,
1995). Baum and Berta find their behaviorally simulated population’s speed of
overall learning (using numbers of learning events in specific time periods)
affected their population average performance as predicted. Population
performance suffered in very speedy and very slow learning populations. The
authors argue that some populations learn too slowly to harvest learning while
others learn too rapidly to discover effective routines.
Mezias and Eisner’s discussion of varied modes of interorganizational
learning also explores the impact of speed in population level transformation. Mezias and Eisner argue outcome based imitation will result
in the most rapid diffusion of a new high performance organizational form,
while outcome imitation will produce more frequent change in the mix of forms,
but the magnitude of such changes will be more modest. Since these learning
dynamics may themselves affect population/industry survival in the face of
competing populations/industries, imitation mode may affect population survival
or prosperity indirectly through its impact on the pace of population level
learning.
These and other
papers taken together imply the relative importance of behavioral versus
cognitive learning may moderate the impact of both speed and variance
(including contrasts between failure and success) in the impact of different
learning levels on population level performance. If other organizations can
easily generate valid insights and strategies by observing failing
organizations, including those that fail from overly rapid organizational
learning, the population/industry may thrive in the presence of high failure
rates or overly-rapid learning by individual organizations. To the degree such
population/industry such inferential learning is difficult, dangers of overly
rapid organizational learning increase.
THE FUTURE
OF POPULATION LEVEL LEARNING RESEARCH
The population
level learning framework represents work in progress. Although learning models
are attractive because they embrace dynamic processes and fit contemporary
emphasis on the role of knowledge, their very richness and complexity is
something of a disadvantage in focusing further research. By raising several
common themes in different contexts, the studies in this volume help offer a
focus for further investigation, four of which we emphasize here.
More systematic
empirical research on how features of neighborhoods of interaction
moderate learning’s impact on population/industry heterogeneity, survival and
prosperity shows promise. It may make sense to contrast learning models here
with related work examining regional populations and more traditional
agglomeration processes (Krugman, 1995). Similarly,
empirical research on the impact of different interorganizational
learning modes will be important. Careful attention to existing formal
models of interorganizational imitation unarguably
implies that interorganizational learning processes
can produce heterogeneity (Miner & Raghavan,
1999). Because most prior empirical research has emphasized only one learning
mode, we have little systematic evidence about their actual relative impact on
industry and population outcomes and implications for strategic management.
One important
notion presented in this volume is that certain population level contours,
such as shared performance standards or even identities, can arise from
learning processes, but then also shape future behavior, including future
learning at lower levels. This important insight implies that potentially
unanticipated collective outcomes may limit future strategic choice for
learning organizations in unexpected ways. Systematic empirical research on
such processes may reveal that they are much more than simple moderators
influencing the impact of repeated organizational or interorganizational
learning. Research on the competitive structures in which learning outcomes are
judged may also have important strategic implications (March, 1991).
Deeper exploration of interactive and situated learning, as well as
the notion of networks as the memory banks for populations or industries
deserve continued attention (Koput, Smitth-Doerr & Powell, 1997; Suchman,
1995). In fact, the broad definition of learning used in the population level
learning framework mandates careful empirical work to tease out what specific
learning processes apply under what conditions This emphasis
is consistent with growing empirical work on specific learning processes (Argote, 1999). The notion that networks may represent some
form of memory – a learning element – for example,
represents an important focus for continued empirical work. It also makes sense
to examine related analytic work done by evolutionary economists and historians
who have considered learning models in some detail (England, 1994).
Other important
issues for continuing work were not highlighted in this volume. For example,
related prior research indicates the level of collective experience can
affect organizational survival chances, which in turn influence
population/industry survival and prosperity. (Baum &
Ingram, 1998). Research on ship-building and hotels suggested individual
organizations may not be able to continue learning from others after their
founding (Argote,Beckman,
& Epple, 1990; Baum & Ingram, 1998). However,
the potential impact of what can be learned after start-up is relatively weak
compared to irreversible initial design and strategy choices because of the
large capital investments needed in these industries. Future work should
examine conditions under which ongoing population experience may enhance the
survival and growth of current firms.
Additionally,
clearly, some organizations are better at learning than others are (Carley, this volume; Leonard-Barton, 1995; Nonaka & Takeuchi, 1995). Improving the organization’s
ability to learn, from its own experience and that of others has been a
principal focus of the burgeoning literature on organizational learning and
knowledge management (e.g., Grant, 1996). Further work exploring firm level
differences in capabilities in learning from others represents an obvious
focus for future research, going beyond the notion of absorptive capacity to a
more comprehensive vision of strategic choices in targets, modes and
capabilities for interorganizational learning.
The
population-level learning framework suggests that in addition to studying the
internal mechanisms through which firms generate or absorb routines, we should understand
the context within which they learn. For example, Wade and Porac describe
how transferring knowledge from managers who bring core-firm experience to a
firm operating in the periphery can be helpful or not, depending on how
industry evolution alters the opportunity structure, opening or blocking career
paths within core enterprises. Similarly, Carpenter and Westpal
(1999) showed how the influence of board interlocks on firm strategic decision
making depended on the external, social-structural context in which the
directors were embedded. More precise understanding of these areas would have
value not only for firm level strategists, but also for collectivities of firms
and groups such as policy makers charged with the well-being of organizational
populations or industries rather than individual organizations.
International and
historical research by Kogut (1993) and others
(Nelson, 1993) pointed to a final area we think promising for continued research:
the dynamics of populations, regions and industries learning from each other.
In many cases, such learning may involve the transfer of population-level
routines including not only technological standards, for example, but also
routines or strategies for coordinating members of a population. Additionally,
while we have emphasized industries as important populations, populations
defined by geography and culture represent an important frontier. Waller,
Gibson and Carpenter (1999) for example, noted that cultural differences were
likely to influence the type, amount and pace of knowledge transfer at both
firm and population levels. Insight into these processes can inform strategists
concerned with the survival or growth of their own industries or regions, and
appear increasingly relevant in contemporary economic settings.
THEORETICAL
VALUE
We think that a
focus on how recurrent learning influences population and industry
transformation makes a distinct contribution to scholarship in both strategic
management and organization theory.
For strategic
management theory, this work – as we note in several places above – supplements
the existing focus on interactions between learning and competition from the
firm’s point of view. It highlights the potential importance of firms trying to
understand the nonlinear dynamics of the way their own learning may alter the
population/industry context in which they act and continue to learn. It is
consistent with efforts to think more carefully about how regions, nations or industries
can promote their own survival and prosperity in the face of knowledge creation
and learning by competing regions, nations and industries (Kogut,
1997; Porter, 1990; Teece, 1986;).
This work also
continues to develop themes raised by the resource-based view of the firm,
through emphasizing the importance of understanding mixes resources in the form
of routines, strategies and practices at a higher level of analysis. The
framework also opens the door to conceptions of industry development that do not
take the most crucial unit of analysis as the firm itself. To the degree that
it emphasizes organizations, it focuses attention on learning differences among
firms rather than structural regularities. This slight adjustment of focus may
offer insights we could not easily achieve when we assume the nature and
distribution of firms represent the most sensible unit for thinking about
population/industry prosperity. This emphasis may also be timely. Attention to
changes in the mix of routines enacted in a population directs attention to
changes in what is actually getting done rather than to changes in
organizational size or boundaries. To the degree that organizations are
increasingly transient or maintain very cloudy borders, this focus may be
increasingly relevant.
In organization
theory, the population level learning framework seems a natural extension of
learning theories, which have moved from individuals, to groups, to whole
organizations but stopped short of any higher level. By examining the impact of
learning on higher-level collectivity the population level learning framework
responds to Stern and Barley’s call to address organization theory’s ‘neglected
mandate:’ of examining the impact of organizations on the broad social systems
in which they [are] embedded (Stern & Barley, 1996).
While overlapping
in some areas, our framework differs in emphasis from several related
perspectives, however. Research on the diffusion of innovations, for example,
tends to study the fate of individual innovations over time rather than the
overall mix of practices. It could not easily account for Anderson’s case of a
single routine dominating, receding, and returning to prominence in a
population, nor can it easily consider how routines interact with each other.
The population
learning framework also provides a different account of many processes than does neoinstitutional theory. The
population level learning perspective predicts the presence of other
organizations early in an emerging industry may enhance survival of new firms
and industry growth because real discovery of useful knowledge is occurring and
advancing through interorganizational learning. This
is a different claim for why increased density may enhance survival than
arguments based on social legitimacy (Aldrich, 1999; Delacroix & Rao, 1994; Miner & Haunschild,
1995).
This framework
also implies that organizations may converge on some traits while diverging on
others, even without the influence of social legitimacy questions, due to
variation in learning modes targets, and neighborhoods. The framework also
differs from organizational ecology approaches to industry transformation by
shifting the focus of attention to changes in distributions of routines,
strategies and practices in a population rather than to distributions of
organizational forms.
Learning processes
that involve pure behavioral trial and error learning can often be
conceptualized as evolutionary processes at a lower level of analysis and can
be a matter of taste which metaphor is used. Some conceptualizations of nested
evolutionary systems, for example (e.g., Singh & Baum, 1994) can be seen as
a type of learning at a higher level of analysis, but the population learning
terminology may not be needed for additional insight. To the degree transformations
involve aspiration levels, deliberate experimentation, knowledge structures and
cognition, however, learning models contain elements
not easily incorporated in traditional evolutionary models of change. These
models often emphasize blind variation rather than search, and do not envision
a separate system of representations of the world as change goes on. Cultural
evolutionary theorists (Boyd & Richerson, 1985)
offer approaches closer to our framework, but have tended to emphasize
individual level processes. Our perspective predicts organizations and
industries may sometimes achieve useful and valid mental representations of the
world. Yet by assuming barriers, stages and limits to learning the framework
also makes different predictions than do economic
models that assume efficient knowledge spillover.
CONCLUSION
Vital strategy
research has long focused on how firms can balance knowledge creation and
deployment with their efforts to find sustainable competitive advantage (Boisot, Griffiths & Moles, 1997; Teece,
1986; Teece, Pisano & Shuen,
1997). Organizations seek to gain knowledge efficiently and quickly from
others, yet be able to harvest the value of their own knowledge and
competencies without their appropriation by others. Difficulties in accomplishing
this, combined with increasing rates of change in contexts of competition and
technology have persuaded some that learning itself represents a crucial
organizational competency (e.g., Lei, Hitt & Bettis, 1996; Teece, Pisano &
Shuen, 1997). At the same time, international
competition and the pace of technological change have brought to light ways in
which entire populations or industries compete with each other over time (Kogut, 1993; Porter, 1990).
The papers in this
volume explore how parallel independent organizational learning, recurrent interorganizational learning and collective population
learning processes can change in the nature and mix of routines, strategies and
practices in a population of organizations (or industry). The population level
learning perspective supplements existing work with two distinct emphases.
First, it directs attention to the impact of recurrent learning at different
levels on population level transformation, rather than only on the firms
directly involved in the learning. Second, it emphasizes change in the mixes of
routines/ practices in the population, rather than change in the mix of
organizations themselves. This focuses attention on what the organizations in a
population are actually doing rather than on how many organizations the
population contains, a focus that may be more relevant if boundaries become
more plastic. These emphases lead to models that often underscore multiple
interactions between competition and learning, and suggest important issues
pertinent to strategic managers and policy makers. We invite the reader to
consider the work here, and to join the conversation represented by this rich
set of papers.
REFERENCES
Aldrich, H.E., Organizations
evolving. London: Sage Publications.
Aldrich, H.E.,
Zimmer, C. R., Staber, U. H., Beggs,
J. J. (1994). Minimalism, mutualism, and maturity: The evolution of the
American trade association population in the twentieth century. In J. Buam & J. Singh (Eds.), Evolutionary
dynamics of organizations (pp. 223-39). New York: Oxford University
Press.
Aldrich, H. E., & Sasaki, T. (1995). R&D consortia in the United States and Japan. Research
Policy¸ 24, 301-316.
Antonelli, C. (1990). Profitability and imitation in the
diffusion of process innovations. Rivista
Internazionale di Scienze Economiche e Commerciali, 37, 109-126
Argote, L. (1999). Organizational Learning: Creating, Retaining &
Transferring Knowledge. Boston: Kluwer.
Argote, L., Beckman, S. L., & Epple, D. (1990). The persistence and
transfer of learning in industrial settings. Management Science, 36, 140-154.
Argote, L., & Epple,
D. (1990). Learning curves in manufacturing. Science, 247, 920-924.
Bateson, G.
(1972). Steps to an ecology of mind; collected
essays in anthropology, psychiatry, evolution, and epistemology. San
Francisco: Chandler Publishing Company.
Baum, J.A.C.
(1996). Organizational ecology. In S. R. Clegg, C.
Hardy, & W. R. Nord (Eds.), Handbook of Organization Studies (pp.
77-114). London: Sage.
Baum, J.A.C., & Ingram, P. (1998). Survival-enhancing learning in the Manhattan Hotel Industry,
1898-1980. Management Science, 44, 996-1016.
Boisot, M., Griffiths, D., & Moles, V.
(1997). The dilemma of competence: differentiation versus integration in the
pursuit of learning. In R. Sanchez & A. Heene (Eds.), Strategic Learning and Knowledge
Management (pp. 65-82). Chichester: John Wiley & Sons.
Burgelman, R. A. (1988). Strategy making as a social learning process: the case
of internal corporate venturing. Interfaces, 18, 74-85.
Cangelosi, V. E., & Dill, W. R. (1965). Organizational
learning: Observations toward a theory. Administrative Science Quarterly, 10,
175-203.
Carpenter, M.,
& Westphal, M. J. (1999). A network perspective
on how outside directors impact strategic decision making. Academy
of Management Best Papers Proceedings.
Corey, E. R.
(1997). Technology fountainheads: The management challenge of R&D
consortia. Boston. Harvard Business School Press.
Cowan, G. A.
(1994). Conference opening remarks. In G. A. Cowan, D.
Pines, & D. Meltzer (Eds.), Complexity: Metaphors, models, reality
(pp. 1-4). Reading, MA: Addison-Wesley.
Cyert, R. M., & March, J. G. (1963).[1992] A
behavioral theory of the firm. Englewood Cliffs, NJ: Prentice-Hall.
Daft, R. L., & Huber G. P. (1987). How organizations
learn: A communication framework. Research in the sociology of organizations,
5, 1-36. Greenwich, CT: JAI Press.
Darr, E. D., Argote,
L., & Epple, D. (1995). The acquisition,
transfer, and depreciation of knowledge in service organizations: Productivity
in franchises. Management Science, 41, 1750-1762.
Delacroix, J., & Hayagreeva R. (1994). Externalities and
ecological theory: Unbundling density dependence. In J. A. C.
Baum & J. V. Singh (Eds.), Evolutionary dynamics of organizations
(pp. 255-268). New York: Oxford University Press.
England, R. W.
(1994). Evolutionary concepts in economics. Ann
Arbor: University of Michigan Press.
Epple, D., Argote,
L., & Devadas, R. (1991). Organizational
learning curves: A method for investigating intra-plant transfer of knowledge
acquired through learning by doing. Organization Science, 2, 58-70.
Grant, R. M.
(1996). Toward a knowledge-based theory of the firm. Strategic
Management Journal, 17, 109-122
Hamel, G. (1991). Competition for competence and interpartner
learning within international strategic alliances. Strategic
Management Journal 12, 83-103.
Haunschild, P. R., & Miner, A. S. (1997). Modes of
imitation: The effects of outcome salience and uncertainty. Administrative
Science Quarterly, 42, 472-500.
Holland, J.
(1996). Hidden order: How adaptation builds complexity. Reading, MA:
Addison-Wesley.
Huberman, B. (1989). The adaptation of complex systems.
In B. Goodwin & P. Saunders (Eds.), Theoretical
biology (pp. 124-33). Edinburgh: Edinburgh University Press.
Huff, A. S. (1990). Mapping strategic thought. Chichester:
John Wiley.
Iansiti, M., & Clark, K. B. (1994). Integration and
dynamic capability: Evidence from product development in automobiles and
mainframe computers. Industrial and Corporate Change, 3, 557-605.
Ingram, P. (1997).
Opportunity and constraint: organizations’ learning from the operating and
competitive experience of industries. Strategic Management Journal, 18,
75-98.
Ingram, P., & Baum, J. A. C. (1997). Chain affiliation and the failure of Manhattan hotels, 1898-1980.
Administrative Science Quarterly, 42, 68-102.
Kauffman, S. A.
(1993). The origins of order: Self organization and selection in evolution.
New York: Oxford University Press.
Powell, W. W., Koput, K. W., & Smith-Doerr, L. (1996). Interorganizational
collaboration and the locus of innovation: Networks of learning in
biotechnology. Administrative Science Quarterly, 41, 116-145.
Krugman, P. (1995). Development, geography, and economic
theory. Cambridge, MA: MIT Press.
Lant, T. K. (1992). Aspiration level adaptation: An empirical exploration. Management
Science, 38, 623-644.
Lant, T. K., & Mezias,
S. J. (1992). An organizational learning model of convergence and
reorientation. Organization Science, 3, 47-71.
Levinthal, D. A., & March, J. G. (1981). A model of
adaptive organizational search. Journal of Economic Behavior and
Organization 2, 307-333.
Levinthal, D. A., & March, J. G. (1993). The myopia of
learning. Strategic Management Journal, 14, 95-112.
Lei, D., Hitt M., & Bettis,
R. (1996). Dynamic core competences through meta-learning and
strategic context. Journal of Management, 22, 549-569
Leonard-Barton, D.
(1995). Wellsprings of knowledge. Boston:
Harvard Business School Press.
Levitt, B., & March, J. G. (1988). Organizational learning. Annual Review of Sociology, 14,
319-340.
Lounamaa, P. H., & March, J. G. (1987). Adaptive
coordination of a learning team. Management Science, 33, 107-123.
March, J. G. (1981). Footnotes to organizational change. Administrative
Science Quarterly, 26, 563-577.
Miner, A. S. (1990). Structural evolution through idiosyncratic jobs: The potential for
unplanned learning. Organization Science, 1, 195-210.
Miner, A. S., & Haunschild, P. R. (1995). Population level
learning. Research in Organizational Behavior, 17, 115-166.
Miner, A. S., & Mezias, S. J. (1996). Ugly duckling no
more: Pasts and futures of organizational learning research. Organization
Science, 7, 88-99.
Mintzberg, H., Ahlstrand,
B., & Lampel, J. (1998). Strategy safari :
A guided tour through the wilds of strategic management. New York: Free
Press.
Mintzberg, H., & McHugh, A. (1985). Strategy formation in an adhocracy. Administrative
Science Quarterly, 30, 160-197.
Mowery, D. C., Oxley, J. E., & Silverman, B. S. (1996). Strategic
alliances and interfirm knowledge transfer. Strategic
Management Journal, 17, 77-91.
Nelson, R. R.
(1993). National innovation systems. New York:Oxford University Press.
Nelson, R. R.,
& Winter, S. G. (1982). An
evolutionary theory of economic change. Cambridge, MA: Belknap Press
of Harvard University Press.
Noda, T., & Bower, J. L. (1996). Strategy making
as iterated processes of resource allocation. Strategic Management Journal,
17, 159-192.
Nonaka, I., & Takeuchi, H. (1995). The knowledge-creating company. New York: Oxford
University Press.
Oster, S. (1999). Modern competitive analysis (3rd edition).
New York: Oxford University Press.
Porac, J. F., Thomas, H., & Baden-Fuller, C., (1989). Competitive
groups as cognitive communities: The case of Scottish knitwear manufacturers. Journal
of Management Studies, 26, 397-416.
Porter, M. E.
(1980). Competitive strategy : techniques for
analyzing industries and competitors. New York: Free Press.
Porter, M. E.
(1990). The Competitive Advantage of Nations and Their
Firms. New York: Free Press.
Quinn, J. B.
(1980). Strategies for change: logical incrementalism.
Homewood, IL: Richard D. Irwin.
Reger, R. K. (1990). Managerial thought structures and competitive
positioning. In A. S. Huff (Ed.), Mapping Strategic Thought (pp. 71-88).
Chichester: John Wiley.
Simon, H. A.
(1996). The sciences of the artificial (3rd
edition). Cambridge, MA: MIT Press.
Smith, F. L., Wright, J. W., & Ostroff, D.
H. (1998). Perspectives on radio and television.
Mahwah, NJ: Lawrence Erlbaum.
Spender, J. C.
(1989). Industry recipes:The
nature and sources of managerial judgement. Oxford:
Blackwell.
Stern, R. N.,
& Barley, S. R. Organizations and social systems: Organiztion
theory’s neglected mandate. Administrative Science Quarterly, 41, 141-146.
Teece, D. J. (1986). Profiting from technological innovation: implications for
integration, collaboration, licensing and public policy. Research
Policy, 15, 285-305
Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic management. Strategic
Management Journal, 18, 509-533.
Van Valen, L. (1973). A new evolutionary law.
Evolutionary Theory, 1, 1-30.
Waller, G., & Carpenter, M. (1999). Time’s arrow: The
impact of differences in the time perspective of knowledge management in a
multicultural context. Presentation, Academy of Management,
Chicago, IL.
Walsh, J. P.
(1995). Managerial and organizational cognition: Notes from a trip down memory
lane. Organization Science, 6, 280-321.
Walsh, J. P., & Ungson, G. R. (1991). Organizational memory. Academy of Management Review, 16,
57-91.
Weick, K. E. (1979). The social psychology of organizing (2nd
ed.). Reading, MA: Addison-Wesley.
Wernerfelt, B. (1984). A resource-based view of the firm.
Strategic Management Journal, 5, 171-180.
Winter, S. G. (1987). Knowledge and competence as strategic assets.
In D. J. Teece (Ed.), The
competitive challenge: strategies for industrial innovation and renewal (pp. 159-183).
Cambridge, MA: Ballinger.