You are on page 1of 14


Manage. Decis. Econ. 29: 489–502 (2008)

Published online in Wiley InterScience
( DOI: 10.1002/mde.1413

Building Micro-foundations for the

Routines, Capabilities, and
Performance Links
Peter Abella, Teppo Felinb, and Nicolai Fossc,*
London School of Economics, London, UK
Marriott School of Management, Brigham Young University, Provo, UT, USA
Copenhagen Business School, Center for Strategic Management and Globalization,
Porcelainshaven 24, Frederiksberg, Denmark

Micro-foundations have become an important emerging theme in strategic management. This

paper addresses micro-foundations in two related ways. First, we argue that the kind of macro
(or ‘collectivist’) explanation that is presently utilized in the capabilities view in strategic
management}which implies a neglect of micro-foundations}is incomplete. There are no
mechanisms that work solely on the macro-level, directly connecting routines and capabilities
to firm-level outcomes. While routines and capabilities are useful shorthand for complicated
patterns of individual action and interaction, ultimately they are best understood at the micro-
level. Second, we provide a formal model that shows precisely why macro-explanation is
incomplete and which exemplifies how explicit micro-foundations may be built for notions of
routines and capabilities and how these impact firm performance. Copyright # 2008 John
Wiley & Sons, Ltd.

INTRODUCTION located at the ‘micro-level,’ that is, the level of

individual action and (strategic) interaction. It
Micro-foundations have become an important seems that strategic management is now embark-
emerging theme in strategic management. Scholars ing on a micro-foundations project somewhat
increasingly realize that understanding such issues similar to similar projects in (macro)economics
as value appropriation (Coff, 1999; Barney, 2001; (Leijonhufvud, 1968; Lucas, 1977) and rational
Lippman and Rumelt, 2003a), resource value choice sociology (Elster, 1989; Coleman, 1990;
(Lippman and Rumelt, 2003b; Foss and Foss, Abell, 2003a, b).
2005), strategy implementation (Barney, 2001), This paper contributes to the emerging micro-
factor market dynamics (Makadok and Barney, foundations project theoretically and methodolo-
2001), inertia (Kaplan and Henderson, 2005), and gically. Specifically, we address the emphasis
firm-level heterogeneity (Gavetti, 2005; Felin and placed upon routines and capabilities as key
Hesterly, 2007) requires that substantial attention constructs in much of strategic management
be paid to explanatory mechanisms that are research and question the present emphasis on
collective and macro constructs. A central argu-
ment in much work in strategic management is
*Correspondence to: Copenhagen Business School, Center for
Strategic Management and Globalization, Porcelainshaven 24, that routines or capabilities are fundamental units
Frederiksberg, Denmark. E-mail: of analysis, and that organizations should be

Copyright # 2008 John Wiley & Sons, Ltd.


conceptualized as repositories of routines and of furthering the received capabilities view in

capabilities (e.g., Nelson and Winter, 1982; Kogut strategic management (cf. Zollo and Winter,
and Zander, 1992). It is, furthermore, asserted in 2002). Specifically, we argue that the nature of
this stream of research that routines and capabil- routines is to internalize externalities. This argu-
ities cause firm-level outcomes, such as financial ment harmonizes with the emphasis in the
performance, innovation, and the boundaries of literature on routines as coordinating devices
the firm (e.g., Nelson and Winter, 1982; Kogut and (Nelson and Winter, 1982, Chapter 5). However,
Zander, 1992; Teece et al., 1997; Eisenhardt and because of asymmetric information routines only
Martin, 2000; Winter, 2003). Thus, it is argued imperfectly internalize externalities. This second-
that explaining firm-level outcomes should take best argument harmonizes with the emphasis in
place in terms of other firm-level variables. the literature which suggests that routines are
Two explanatory gaps}of which we shall often not optimal (Nelson and Winter, 1982).
primarily concentrate on the second one}stand However, we diverge from the literature, first, by
out in this research stream. First, there is little explicitly modelling the micro-foundations of
recognition of the need to explain the origins (or how routines impact performance; second, by
emergence) of routines and capabilities (except embedding our arguments in a conventional
perhaps in terms of other routines and capabil- production function framework; and third, by
ities). Second, exactly how routines and capabil- modelling production externalities as giving rise to
ities are related to firm-level outcomes, such as prisoner’s dilemma situations rather than to
performance, is seldom elaborated (cf. Argote and coordination problems. Finally, we link routines
Ingram, 2000, p. 156). Thus, crucial explanatory and capabilities in a simple manner by arguing
theoretical mechanisms are left unexplored and that a firm can be described as possessing the
implicit. capability to realize a routine to the degree that it
We argue that gaps related to underlying micro- can repeatedly internalize such externalities (i.e.
foundations cannot be bypassed, they need to be realize synergies).
explicated, and that in addressing these gaps one
must involve the level of individual action and
interaction. The specific gap and associated
problem is that the macro (or ‘collectivist’) mode ANALYTICAL LEVELS IN STRATEGIC
of explanation that currently dominates large parts MANAGEMENT
of the strategic management literature, and which
asserts a causal relation running directly from Many phenomena of interest in the strategic
routines and capabilities to firm-level outcomes, is management field, such as financial performance,
incomplete. To be sure, firm-level concepts such as diversification patterns, vertical integration, com-
routines and capabilities may be (indeed, are) petitive rivalry, etc., are placed on a level of
relevant to the explanation of firm-level outcomes. analysis that is above that of the individual. In
However, they are relevant because they are useful fact, explananda (i.e. the dependent variables) in
shorthand for complicated repetitive patterns of strategic management are usually placed at the
individual action and coordinated interaction. level of the firm. However, the explanans (i.e. the
Thus, the micro-level (i.e. individual action and independent variables and the mechanisms that
interaction) ultimately replaces the macro-level link them to the dependent variables) may involve
(i.e. the postulated direct link between routines/ other levels of analysis as well, such as the dyadic
capabilities and performance) in the explanation level, the industry level, or the level of individuals.
of how routines/capabilities and performance are Any theoretical and empirical effort to explain
linked. phenomena in strategic management then has to
To clarify this argument, we develop a formal make a choice that concerns the level(s) at which
model that details the importance of the micro- explanation takes place (Dansereau et al., 1999). A
level in explaining firm-level outcomes. The argu- classic distinction in social science research is
ments, and the accompanying modelling effort, between the collective and the individual level
explain how micro-foundations can be built for (Lazarsfeld and Menzel, 1970; Coleman, 1990,
capabilities and how they are linked to firm-level pp. 3–5), which in the context of organizational
outcomes. Thus, the paper is offered as one way theory and strategic management corresponds to a

Copyright # 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 489–502 (2008)
DOI: 10.1002/mde

distinction between ‘macro’ and ‘micro.’ We argue filled with different theoretical mechanisms,
that strategic management research has too often entirely dependent on theory development on the
located not only the explanandum (which is part of the analyst. (Our later modelling effort is
entirely legitimate) on the collective or macro- an example of development of such concrete
level, but also all of the explanans (which is often theoretical content.)
Macro-Explanation in Strategic Management
At first inspection, the framework depicted in
A General Model of Social Science Explanation
Figure 1 would seem to formally allow for
In order to clarify notions of ‘micro’/‘individual explanation that takes place solely in terms of
level,’ and ‘macro’/‘collective level’, as well as arrow 4, that is, explanatory accounts that are
examine the relations between these notions and wholly located on the macro-level. However,
levels, consider Figure 1 which builds on the whether arrow 4 explanation is deemed legitimate
framework popularized by James Coleman (1990). depends on (ontological) criteria related to an
This framework organizes much of our discussion understanding of how the social world works
and modelling effort. (Mäki, 2001). Specifically, there are no conceivable
The Figure makes a distinction between the causal mechanisms in the social world that operate
macro-level and the micro-level. For example, it solely on the macro-level. There are no macro-level
may be that the macro-level is organizational entities on the social domain that somehow possess
whilst the micro-level is that of individuals. As capacities or dispositions to act (Cartwright, 1989)
shown, there are links between macro–macro that make them capable of directly producing
(arrow 4) and macro–micro (arrow 1), micro– macro-level outcomes, and there are no processes
micro (arrow 2), and micro–macro (arrow 3).1 The of interaction between macro-entities that take
Figure also makes a distinction, perhaps place on this level. In short, there is no macro-level
more implicit, between what is to be explained causal mechanism that can be theoretically repre-
(i.e. the explanandum) and its explanation (i.e. the sented in terms of arrow 4.2
explanans). In social science, the aim usually is to However, arrow 4 explanation is not necessarily
explain either a macro-level phenomenon (located entirely ruled out. First, arrow 4 may be taken as no
in the upper right hand corner of Figure 1), such as more than a representation of a correlation between
a firm-level outcome, or a link between macro- macro-variables in need of further explanation of
phenomena, as indicated by arrow 4. An example the micro-level. This is entirely unproblematic.
of the latter may be an observed correlation Second, arrow 4 may be used as convenient
between the routines and the performance of firms shorthand. ‘Convenient shorthand’ here means that
in a population. To explain and understand a we can make use of arrow 4 explanations when we
particular phenomenon (such as overall firm are convinced that they can be reduced to micro-
performance) the analyst makes use of theoretical mechanisms, but performing this reduction would
mechanisms that are consistent with the arrows. not add anything in the explanatory context (cf.
Note that the arrows in Figure 1 are, from a Stinchcombe, 1991). For example, there is no
theoretical perspective, empty boxes. They may be problem in asserting and showing that organiza-
tional culture perhaps is correlated with organiza-
tional performance. More generally, arrow 4
”Social 4 Social
facts”(e.g., outcomes explanation may be legitimate when the relation-
institutions) • •
”macro” ship does not appear to be particularly puzzling,
for example, because we have a good grasp of the
1 3 underlying micro-mechanism (Abell, 2003b,
p. 261).3 Be that as it may, it certainly is the case
that several examples of arrow 4 ‘explanation’ can
Conditions • Individual
be found, such as the arguments that routines are a

of Individual
2 action direct cause of firm-level adaptation (Nelson and
Winter, 1982), ‘combinative capabilities’ cause firm-
Figure 1. A general model of social science explanation. level innovativeness (Kogut and Zander, 1992), and

Copyright # 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 489–502 (2008)
DOI: 10.1002/mde

different ‘absorptive capacities’ cause differences in other stakeholders (Coff, 1999; Lippman and
how well firms learn from partner firms in inter- Rumelt, 2003a, b). In a related vein, to say that
organizational relations (Lane and Lubatkin, 1998). a firm has a certain capability is essentially
It was briefly suggested above that macro- shorthand for a complex set of underlying
explanation may be warranted under certain individual actions and interactions, and associated
conditions. However, it would be hard to argue characteristics or skills which make the realization
that these conditions are always met in strategic of these capabilities possible. Because scholars
management research; in fact, it is rarely so. Thus, may not always want to make explicit reference to
strategic management scholars do not have the- complicated underlying patterns of actions, they
ories of why routines and capabilities impact firm often prefer to make use of explanatory shorthand
performance that involve the micro-level, that is, in the form of collective concepts. This is
at the level of individual action and interaction. completely legitimate. However, a fundamental
Second, there is much reason to think that methodological (and ultimately theoretical and
micro-level considerations add substantially to managerial) problem in contemporary strategic
macro-level understanding. For example, a micro- management research is that it seems to be too
perspective suggests that macro-level heterogeneity often forgotten that explanation in strategic
can be an epiphenomenon of individual level self- management should nevertheless have a micro-
selection. Further reasons why micro-foundations foundation.
are in fact critical are given in the following. Before proceeding to our modelling effort, we
delineate, building on Coleman’s (1990, pp. 3–4)
insight, perhaps the most persuasive reason for
Why Micro-Foundations Are Critical
why micro-foundations are critical for strategic
We take the position}associated with management. That is, perhaps the most critical
‘methodological individualism’}that the explana- problem with macro-level explanation is that there
tion of firm-level (macro) phenomena in strategic are likely to be many alternative lower-level
management must ultimately be grounded in explanations of macro-level behaviour which
explanatory mechanisms that involve individual cannot be rejected with macro-analysis alone.
action and interaction (cf. Hayek, 1952; Ullmann- Even if a large sample can be constructed on the
Margalit, 1977; Elster, 1989; Coleman, 1990; basis of macro-units of analysis, a problem of
Boudon, 1998). We also take it that the ultimate alternative explanations may persist. As indicated
aim of scientific endeavour in the field of strategic above, alternative explanations at lower levels are
management should be to identify and theorize the readily apparent in, notably, the capabilities view,
causal mechanisms}the ‘cogs and wheels’ (Elster, which seeks the explanation of differential firm
1989, p. 3)}that produce the observed associa- performance in firm-level heterogeneity, that is,
tions between events (Cowan and Rizzo, 1996; heterogeneous routines and capabilities. However,
Hedstrom and Swedberg, 1998).4 heterogeneity may be located at the individual
Combining methodological individualism with level, notably when individuals self-select into
an emphasis on causal mechanisms implies that particular firms.
strategic management should fundamentally be An argument for the importance of under-
concerned about how intentional human action standing micro-foundations lies in the fundamen-
and interaction causally produce strategic phe- tal mandate of strategic management: to enable
nomena. It is implicit in this view that explanatory managers to gain and sustain competitive advan-
black boxes be avoided (Boudon, 1998). Admit- tage. To achieve this, managerial intervention is
tedly, black boxes may sometimes be justified in required, which inevitably has to take place with
terms of explanatory parsimony (Hedström and an eye toward the micro-level.5 Coleman (1990, p.
Swedberg, 1998, p. 12; also see Coleman, 1990, p. 3) convincingly argues that explanations that
16), as indeed happens in much of arrow 4-type involve the micro-level have the properties of
explanation. Strategic management scholars know being more stable, fundamental, and general than
(or should know) that when they speak of a firm macro-level explanations:
appropriating a revenue stream, this is shorthand
for a complicated underlying process of bargaining An explanation based on internal analysis [i.e.
between numerous individual resource owners and micro-foundations] of system [organization]

Copyright # 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 489–502 (2008)
DOI: 10.1002/mde

behaviour in terms of action and orientations of link is merely metaphorical. Routines are con-
lower-level units is likely to be more stable and ceptualized by Nelson and Winter (1982, p. 124) as
general than explanation which remains at the the ‘skills of an organization’ and as ‘a repetitive
system level. Since the system’s behaviour is in pattern of activity in an entire organization’ (1982,
fact resultant of the actions of its component p. 97). Routines refer to repetitive interaction that
parts, knowledge of how the actions of these is somehow patterned, typically (but not necessa-
parts combine to produce systematic behaviour rily) in the form of fixed sequences of individual
can be expected to give greater predictability actions where the specific sequence and the
than will statistical relations of surface char- contents thereof are organization-specific (i.e. firm
acteristics of the system. A may do things in a different order than firm B)
(Cohen et al., 1996; Dosi et al., 1999; Becker,
To the extent that strategic management is 2004).
concerned not just with explaining past perfor- While Nelson and Winter spend considerable
mance but also with being prescriptive, Coleman’s time on developing the notion of a routine, they
point raises an important concern: the ability to are less forthcoming about the notion of cap-
predict is a condition for putting forward pre- ability, which is loosely defined as ‘associated with’
scriptions. Micro-foundations are therefore an ‘individual members’ repertoires . . . particular
important part of strategic management as a collections of specialized plant and equipment . . .
prescriptive enterprise. [and]. . . the ability to operate that plant and
equipment’ (Nelson and Winter, 1982, p. 103).
Unfortunately, they do not clarify how routines
and capabilities are related, and much the same
ROUTINES AND CAPABILITIES may be said of the subsequent literature that has
taken its cues from Nelson and Winter.8 Because
The seminal and in many ways founding contribu- of this lack of clarity with respect to the capability
tion to the capabilities view is Nelson and Winter construct, we shall primarily make reference to the
(1982). Their conceptualizations and insights have less ambiguous routines construct, but later
been fundamental to the way subsequent work on suggest a specific interpretation of what a cap-
routines and capabilities has developed (Foss, ability may entail.
2003; Becker, 2004), not the least in strategic Neither Nelson and Winter, nor subsequent
management (e.g., Teece et al., 1997; Dosi et al., writers in strategic management, have (to our
2000; Eisenhardt and Martin, 2000).6 In this knowledge) offered a rigorous analysis of why and
section, we look at routines, first, as dependent how actions taken by different individuals in an
variables (i.e. as explananda), and, second, as organizational setting should come to mesh into
independent variables (i.e. part of the explanans). orderly and repetitive (reproducible) sequences
We argue that in both cases the extant literature (employee A doing X after employee B has done
has a problem with missing micro-foundations. Y, etc.), that is, routines. It is arguable that the
reason for our understanding of routines being
incomplete in this manner is the lack of an explicit
Explaining Routines
starting point in individual action and interaction.
Nelson and Winter begin their analysis of routines Specifically, it is necessary to examine the actions
from the notion of skill (Nelson and Winter 1982, that an individual can take (e.g., routine action or
Chapter 4), which they define as ‘. . . a capability non-routine action) and the payoffs associated
for a smooth sequence of coordinated behavior with these actions before it is possible to ascertain
that is ordinarily effective relative to its objectives, whether the actions individually taken will con-
given the context in which it normally occurs’ stitute a routine.
(1982, p. 73).7 There are a number of reasons why Consider the left-hand side of the diagram
the skill metaphor is attractive to Nelson and depicted in Figure 2, which is concerned with the
Winter (see Foss, 2003), but the one that is of explanation of routines (i.e. routines as explanan-
interest in the present context is that the notion of dum). In terms of the diagram, arrows 1–3 are not
skills is used to establish a link between individual given theoretical content in extant work on
action and organizational routines, even if that routines. Instead, routines at time t1 are explained

Copyright # 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 489–502 (2008)
DOI: 10.1002/mde

Macro antece- Firm-level

dents for routines 4 Routines 4a outcomes
• • •
Not described and Not described and
1 clrified in extant clrified in extant
literature on routines 1a literature on routines 3a

• • • •
2 2a

Figure 2. Explaining routines and explaining by means of routines.

directly in terms of routines at t0. For example, certain routine may be a source of superior
Nelson and Winter (1982) argue that routines performance}requires taking a starting point in
change through the operation of other routines individual action and interaction: the routine may
(‘dynamic routines’). This is an explanation in be associated with a high-productivity equilibrium
terms of arrow 4. A similar neglect of the micro- (Leibenstein, 1987), for example, because it leads
level arises in connection with explanation that to superior coordination of actions (Camerer and
involves routines, not as explananda, but as part of Knez, 1996) (sans incentive conflicts) or because it
the explanans. leads to agents choosing actions that overcome
latent prisoner’s dilemma situations. Thus, the
causal links from routines to firm-level outcomes
Explaining by Means of Routines are never direct (arrow 4a in Figure 1); rather, they
Among the reasons why routines have proved involve individual skills, motivations, and actions.
attractive to strategic management scholars is that Unfortunately, these individual level considera-
they are seen as representing the outcomes at a tions have been consistently blackboxed in the
given time of a firm’s knowledge development path received capabilities view.9 The following section is
(e.g., Teece et al., 1997; Eisenhardt and Martin, an attempt to open up this black box.
2000). They are therefore relevant to the under-
standing of such important knowledge-based
phenomena as heterogeneity, competitive advan- EXPLAINING THE LINKS BETWEEN
tage, inertia, diversification patterns, and patterns ROUTINES, CAPABILITIES, AND
of innovation. However, as Argote and Ingram PERFORMANCE
(2000, p. 156) lamented, to the extent that there
has been progress in studying knowledge as the
Conceptualizing Routines and Capabilities
foundation of competitive advantage, ‘. . . it has
been at the level of identifying consistencies in As noted, the relation between the core constructs
organizations’ knowledge development paths and of routines and capabilities is far from clear in the
almost never at the level of human interactions literature, and definitions of these constructs tend
that are the primary source of knowledge and to be vague. We suggest the following simple
knowledge transfer.’ In other words, explanations definition of routines and capabilities and how
of (for example) competitive advantage that they relate: A firm can be described as possessing
involve notions of routines in the explanans the capability to realise a routine to the degree that
typically reason directly from these to competitive it can repeatedly internalise a pattern of individual
advantage. In terms of the right-hand side of level external productivity effects.10
Figure 2, this amounts to explanation using arrow This definition seems to capture important parts
4(a). Again, however, arrows 1(a), 2(a) and 3(a) of what many scholars}not only Nelson and
are not given theoretical content. Winter}imply by routines and capabilities (e.g.,
Understanding the firm-level consequences of Cyert and March, 1963, pp. 120–133).11 Notably,
actions being routinized}for example, why a there is more to a routine than merely sequentially

Copyright # 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 489–502 (2008)
DOI: 10.1002/mde

organizing the productive effort of a number of conception is consistent with the notion that
independent productive agents. Their efforts are routines assist in coordinating dispersed, tacit
interdependent (as manifested in external produc- knowledge (Kogut and Zander, 1992; Cohen
tivity effects), and these interdependent efforts can et al., 1996; Dosi et al., 1999), that is, knowledge
be repeated (Cohen et al., 1996). Also note the that cannot be fully centralized in the management
cross-level nature of this definition referring, as it team. Routines, whilst conferring potentially high
does, to both a firm (collective level) and financial performance, may not be optimal (Nel-
individuals. son and Winter, 1982, p. 126).14 Finally, the
One might ask why the routine should not be averaging procedure harmonizes with the key idea
attributed to the (collective action of) individuals in the literature that an important function of
rather than to the organization (thus obviating the firms is to simplify the micro-complexity of inter-
need for an awkward cross-level conceptualiza- individual external productivity effects by means
tion). The reason for taking this route is that the of routines and standard operating procedures (cf.
literature appears to make it a requirement of Cyert and March, 1963).15
routines that they are replicable by mechanisms
operating at the organizational level (Nelson and
Winter, 1982, p. 117; Cohen et al., 1996).12 To put Fundamental Notions
it somewhat differently, routines are deemed to be
institutionalized to the extent that they are not The following modelling exercise gives some
overly sensitive to the turnover of employee and substance to the explanatory skeleton represented
management turnover (and perhaps depreciation by the diagram in Figure 1. Specifically, Figure 3,
of substitutable capital assets) in realizing the which is simply an application of the Coleman
capability (Nelson and Winter, 1982). This feature diagram in Figure 1 to the present model,
must, of course, be a matter of degree and it is introduces some of the notation and terminology
difficult to precisely characterize it. used here.
The basic analytical procedure is as follows. N
individuals exerting certain skills at a certain level
Firms as Averaging Mechanisms of motivation, X, could}in the absence of
The way we propose to address these issues is by externalities in production}operate indepen-
conceiving a firm as an averaging mechanism. This dently producing an aggregate output Yindep..
notion may be exemplified in terms of a principal- Under standard assumptions about production
agent setting with one principal and a number of costs (which we shall leave implicit for the sake of
agents that cooperate in a team (as in Alchian and clarity), there is in this case nothing to be gained
Demsetz, 1982). Information is asymmetric in the from routinization. In order to provide a rationale
specific sense that the principal cannot observe for routines, we introduce production externalities
individual efforts and outputs. He can only as a network (i.e. a di-graph). Optimal output,
 ; now requires micro-level internalization of
observe the team’s output; however, basing the
remuneration of individual team members on team these effects. Consistent with the theory of the firm
output introduces a prisoner’s dilemma problem. literature (e.g., Alchian and Demsetz, 1972;
Resorting to some kind of monitoring is therefore
necessary. Although he cannot observe individual
R,Z 4 Y
effort, the principal/manager can, based on var-
• •
ious signals, form an estimate of the average of
input productivities and therefore an estimate of
output, given the average. Moreover, we assume
that managers can implement this average and that 1 1a 3
the means to such implementation is a routine.13
The average mechanism conception implies that
firms (i.e. management) do(es) not have the
information to internalize the full micro-complex- X • • Individual
ity of external effects, which is why resorting to
some averaging procedure is necessary. This Figure 3. Explaining the routines/performance link.

Copyright # 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 489–502 (2008)
DOI: 10.1002/mde

Holmström, 1982; Grossman and Hart, 1986) we productive factors (notably capital) which may be
assume that it is beyond the individuals, acting regarded as embodied in b0.
independently, to achieve such internalization.
In this context, the firm (i.e. management) may
be seen as a mechanism for attending to the Aggregating Up
external effects when the output is Ydep. Because of Under standard assumptions about the value
asymmetrical information, not all external effects (benefits) of Y and the cost of motivated skill (X)
can be (efficiently) internalized; hence, the to each individual, the optimal levels of Yi, i ¼
notion that only average external effects are 1; 2; . . . ; i; . . . ; N; are easily definable in terms of
internalized. The application of a routine results equalizing costs and benefits at the margin. Then
in an output level that lies somewhere in the the total output, Yindep., is given by
interval, ]Yindep., Ydep:
Yindep: ¼ Si Yi ; ð3Þ
In the following, routine impact on firm
performance is defined as the explanandum, that where ‘indep.’ stands for independent individual
is, we primarily look at the right-hand side of the maximization. Thus, the firm-level outcome is
diagram in Figure 2 (and black-boxing the left- reached by simple addition (i.e. arrow 3 in Figure
hand side). If a firm is conceived as a repository of 2). However, the notion of routines implies more
routines, they are not necessarily independent of than Equation (3). In order to better capture the
one another, as, reflected in notions of routine meaning and implications of routines, assume now
hierarchies (Nelson and Winter, 1982; Winter, that the individual production functions poten-
2003). The capability to realize one routine may tially take the form
depend upon the capacity to realize other routines
Y  ¼ b ðX S a X Þb1 e ;
i 0 i k ik k i ð4Þ
(i.e. inter-routine synergies). For the sake of
expository convenience, however, we abstract where ei represent stochastic factors. In logs, this
away from such interdependence and deal with becomes
independent routines. log Yi ¼ log b0 þ b1 log Xi þ b1 log Sk ðaik Xk Þ
þ log ei ; ð5Þ

Individual Level Considerations where k ¼ 1; 2; . . . ; N; k=i: ei has the usual

stochastic interpretation across the N individuals.
To introduce individual level considerations, let aik is the weighting of the external effect of
the productive output of individual i ¼ ð1; 2; . . . ; individual k’s motivated skill (Xk) upon individual
NÞ in the routine be Yi. Further, let the exogenous i’s output performance (i.e. ‘Hawthorne effects’).
individual (micro)-level variable be X (i.e. the In other words, aik represents externalities in
bottom node to the left in Figure 2). To ease production. In this context, the notion of
presentation, Xi represents an interactive (choice) ‘institutionalization’ can be interpreted to refer to
variable of individual i’s motivation and skills (i.e. the extent to which the effects represented by aik
‘motivated skill’). More specifically, we can can be maintained in the face of turnover; for
represent an individual level (arrow 2) production example, strong institutionalization implies that
function as a simplified Cobb–Douglas function: aik is rather invariant to personnel turnover.
Yi ¼ b0 Xib1 ri ; ð1Þ It is convenient to interpret the matrix, A, of
binary coefficients aik across the N actors as a
where ri represent stochastic factors. In logs, this network, or more formally, a di-graph, R ¼ ðN; AÞ;
becomes where N represents, the nodes and A the arcs. In
log Yi ¼ log b0 þ b1 log Xi þ log ri ; ð2Þ fact, given our earlier remarks about the institu-
tionalized capacity of routines it may be useful to
ri has the usual stochastic interpretation (i.e. regard the graph R as running across institutiona-
normally distributed with mean zero, uniform lized positions rather than specific individuals.
variance, and zero co-variance among the resi- This conceptualization links capability and rou-
duals) across the N individuals. b0 is the total tines: the more an organization has institutiona-
factor productivity of the routine. Again, to avoid lized such positions, the better its capability of
notational complexities we have suppressed other repeatedly realizing the routine.17 It is as if the

Copyright # 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 489–502 (2008)
DOI: 10.1002/mde

organization has a ready blueprint for organiza- caused by inputs of motivated skill), while Rj is a
tional design (including task allocation) and HRM measure of the extent to which the routine
policy that it can efficiently and repeatedly internalizes externalities. (We will return to this
implement in the face of even substantial personnel below.) By making these stochastic functions we
turnover (cf. Nelson and Winter, 1982). signal that arrow 1 in Figure 3 is empirical rather
The significance of A is that it marks the than definitional.20
potential for collective action in the following Introducing variation across j and combining (2)
sense: to the degree each individual, k, sets her with (8) and (9) we obtain an expression for
level of motivated skill at Xk, taking account of, individual i’s productivity in routine j:
not only her own output Yk, but also the impact
she has on the other individuals in R, the value of log Yij ¼ c01 Zj þ c11 Rj log Xij þ log u0j
þ u1j log Xij þ log rij : ð10Þ
Ydep: ¼ S Y
i i ð6Þ
will be optimal. In general, Note the dependence of the ‘error’ on the value of
X which in the context of empirical estimation
Ydep: > Yindep: ð7Þ
would call for special treatment.
if, for at least one pair i and k, aik > 0; and Variables Z and R are under the control of the
individuals take account of their impact on other firm (i.e. in practice management). Z is any
individuals in R. variable, like size influencing economies of scale,
which impacts total factor productivity. With
respect to Z (our concern here), management has
PRODUCTION IN ROUTINES at its disposal three basic mechanisms for deter-
mining the aggregate input, X (i.e. motivated
If, as before, we assume conventional individual skill)}namely, firstly, an incentive system; second,
cost functions in X, it is intuitive that the monitoring and direct supervision; and, third, the
individual level production functions (i.e. Equa- creation and maintenance of firm-level cultural
tion (4)) establish an N-person prisoner’s dilemma. norms.
Thus, the (Pareto) optimum is achieved when all It is reasonable to assume that management
players internalize their external effects in setting does not possess the detailed information on
their respective X values (Holmström, 1982). Each inputs (X) and the strength of the external effects
has, however, an incentive to free ride and then Y (aikj) in order to design individual-specific incen-
(indep) will be realized, supporting the (sub- tive contracts which would optimally internalize
optimal) N-person Nash equilibrium.18 The firm these effects. Management could offer a collective
tries to prevent this problem by institutionalizing a incentive contract (e.g., profit or gain sharing), but
routine.19 this is open to free-riding (Alchian and Demsetz,
We now allow both parameters in equations of 1972). Indeed, we assume, in the first place, that
the form (4) to vary across routines j. Assume, individual coordination cannot be achieved, thus
without any loss of generalization, that they take, necessitating management (i.e. the firm). Manage-
respectively, the simple linear forms ment will, of course, expend resources on super-
vision and monitoring, but once again the details
log b0j ¼ c01 Zj þ log u0j ; ð8Þ of complex routines will fall beyond their grasp.
Given the potential failure of sharp incentives and
b1j ¼ c11 Rj þ u1j ; ð9Þ monitoring, the establishment of norms (i.e.
where again the u terms are both stochastic with corporate culture) may provide a partial solution
the standard interpretation. Zj and Rj are variables (cf. Miller, 1992). More generally, we conjecture
which vary across routines, but not across that management can only be the recipients of
individuals within a routine. These are firm-level noisy signals about the potential of the routine.
variables that impact/moderate the relation be- In light of this earlier analysis let,
tween individual-level motivated skill and indivi- Rj ¼ 1=NðN  1ÞSi Sk aikj Xik : ð11Þ
dual output performance. Thus, Zj measures the
variation in total factor productivity across That is to say, Rj is the mean value of the
routines (that is, effects in output that are not institutionalized external effects in routine j.

Copyright # 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 489–502 (2008)
DOI: 10.1002/mde

So (10) becomes to the idea whereby arrow 4 can constitute a sui

log Yij ¼ c01 Zj þ c11 ð1=NðN  1ÞSi Sk aijk Xkj Þ log Xij generis form of explanation (Abell, 2003b).
Given this change, (10) also undergoes change,
þ log u0j þ u1j log Xij þ logrij : ð12Þ
namely to
The firm by averaging over the external effects log Yij ¼ c01 Zj þ log rij þ log m0j : ð16Þ
institutionalizes the production functions across
routines: But (16) is depicted as arrow 1a in Figure 3. Given
ð1=NðN1ÞSi Sk aikj Xkj Þ (6), arrow 4 in Figure 3 is a transitive closure of
c c
Yij ¼ Zj 01 ðX Þ
ij 11 u0j rij Xij uij : ð13Þ arrows 1a and 3.
Thus, the above demonstrates that we can use
With an averaging assumption the total output
collective level, arrow 4 explanations at best as
will be Y (mean) for any j, where
shorthand or ‘reduced form’ explanation. Arrow 4
 >Y ð14Þ
Ydep: mean > Yindep: : must always be either a conjunction of mechan-
Notice that the collective/macro-level variables, Zj isms indicated by arrows 1–3 or/and arrows 1a
and Rj, enter the explanation of the routines- and 3. Thus, in explaining collective-level phenom-
performance link, not through arrow 4 in Figure 2, ena, reference must be made to the level of the
but by moderating the relationship between the individual. A further interpretation is that indivi-
exogenous individual-level variable and perfor- dual-level/micro-explanation replaces collective-
mance (c11; arrow 2), or directly by influencing Yij level/macro-explanation.
(i.e. c01 Zj; arrow 1a).
Although the purpose of this paper is rather
general, pointing to the necessity of a micro- CONCLUDING DISCUSSION
foundational model in any theory of routines,
there are some possible empirical tests of our Towards Micro-foundations
particular model. First, insofar as firm perfor-
mance is attributable to institutionalized routines, The field of strategic management seems to be
then collective incentives are likely to be absent. increasingly aware of the need to embark upon a
Second, payment by results is also likely to prove micro-foundations project. There is clearly an
ineffective. Third, for routines to be effective they increasing need to build individual-level founda-
will probably need to be complemented by strong tions for firm-level phenomena, such as hetero-
norms. (We return to empirical issues in the geneity, inertia, and superior financial performance.
Concluding Discussion.) However, the perhaps dominant approach to firm-
level heterogeneity in strategic management, the
capabilities-based view, has seen virtually no
Collectivist Explanation is Incomplete attempts to build explicit micro-foundations. As a
Armed with the above analysis, we can now result, it is unclear how crucial collective or macro-
examine somewhat more rigorously the claim level constructs, such as routines and capabilities,
made earlier that collectivist explanations are, in impact firm-level performance (and it is unclear
the present context, not tenable. Specifically, we how they emerge from individual action and
can ask whether it would ever prove sensible to interaction). This is unsatisfactory from the point
explain the capabilities-performance link only in of view of theory building in strategic management
terms of arrow 4. theory, because crucial underlying mechanisms
Assume (2), introducing variation across j, is remain unspecified. Similarly, the ‘reduced form’
changed to approach of work on routines and capabilities also
means that applied work will suffer from a great
log Yij ¼ log b0j þ log rij : ð15Þ
deal of indeterminacy in the sense that multiple,
This is equivalent to saying that all the individual- potentially rival stories on the micro-level can
level exogenous variables, embodied in rij, bear a explain a macro-correlation.
random relationship to individual performance; Finally, we are worried that the absence of
there is no generalizable impact on performance of micro-foundations in the capabilities view may
any micro-level variables. This is of course highly contribute to a disappearing mandate for strategic
unlikely, but this is the only meaning we can attach management. In other words, the possibility of

Copyright # 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 489–502 (2008)
DOI: 10.1002/mde

strategic action may become obscured by putting explain both the emergence of routines and their
for too much emphasis on firm-level constructs, impact on performance (i.e. the whole of Figure 2).
such as routines and capabilities. It seems critical One reason is that how routines impact perfor-
for both organizational scholars and managers to mance may be related to which routines are
be able to impute action to individual rather than allowed to emerge and such a feedback loop may
collective variables. And, if routines and capabil- be a crucial part of the dynamics of routines
ities indeed are meaningful variables then both (Nelson and Winter, 1982).
strategic management scholars and managers Another limitation resides in our focus on
should take an interest in the micro-level mechan- prisoner’s dilemma games. The usual treatments
isms through which they exert their influence on of routines tend to see them as solving coordina-
firm performance. Performance improvements tion rather than cooperation problems (Nelson
may come about not just through selecting new and Winter, 1982). However, recent work has
routines and capabilities (as in Nelson and Winter, suggested that resolving incentive problems (i.e.
1982), but also through changing or influencing problems of cooperation) may be an important
the micro-mechanisms through which routines part of what routines accomplish (Gavetti, 2005;
work their influence on performance. Kaplan and Henderson, 2005). Our model illus-
The main argument in this paper has, accord- trates exactly this feature of routines. Obviously,
ingly, been that micro-foundations must be built, however, such a treatment leaves out a host of
and a simple formal model has been offered as an other possible aspects of routines (for catalogues
example of how micro-foundations may be built of these, see Levitt and March, 1988; Becker,
for the case of understanding how routines impact 2004). For example, routines may contribute to
performance. In terms of the theories that strategic shaping cognition in a firm. This aspect is left out
management research draws upon, an implication of consideration in the present treatment. Simi-
of this work is that routines can be meaningfully larly, we remain agnostic on the issue of whether
interpreted within a standard production function (or to which extent) routines are emergent or
framework (in contrast, Nelson and Winter, 1982, designed entities (cf. Dosi et al., 1999).
are strongly critical of this framework), and that Developing a clear picture of what routines
key insights from the economic theory of the firm accomplish arguably requires discussing one thing
(Alchian and Demsetz, 1972; Holmström, 1982) at a time. Still, it should be noted that although the
on firms as vehicles for internalizing externalities reasoning in this paper draws on ideas from
may have a significant bearing on the under- economics, it is consistent with a broad set of
standing of the nature of routines. A theoretical behaviours. In particular, the arguments and
contribution that emerges from the latter con- modelling effort in this paper are not tied to
ceptualization is the notion we have proposed of rational choice theory; learning and adaptive
firms as averaging mechanisms. behaviours are entirely consistent with the model.
Given the various, strikingly broad, definitions of
routines (routines as truces, memory etc), of course
Limitations and Future Work
the conceptualization and definition utilized by a
The aim of this research has been to make a particular scholar will influence how the routine-
fundamental methodological point, that collecti- performance link is explained. For example, those
vist or macro-explanation is incomplete, and to who take a more cognitive approach to routines
indicate that it is possible to build micro-founda- and capabilities (e.g., Levitt and March, 1988) have
tions for how routines and capabilities impact suggested an alternative explanation of the routine-
performance. Because of this overall focus, a performance link, along with implications for
number of specific points have (deliberately) not micro-foundations (though these remain implied).21
been developed. For example, we have refrained We welcome such work. In order to make a micro-
from directly discussing the micro-foundations of foundations project viable in such a relatively
routines themselves, and have taken routines to be diverse field as strategic management, alternative
rather deus ex machina like. The present approach micro-foundations should be tried out.
may therefore strike some readers as not going Finally, although the main purposes of this
sufficiently far in the direction of micro-founda- research are methodological and theoretical in
tions. Ultimately, a satisfactory treatment should nature, the issue of how to make micro-foundations

Copyright # 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 489–502 (2008)
DOI: 10.1002/mde

testable and accountable to observed performance hiring key employees (in which case the micro-level
facts must be briefly raised. Recall the definition of is directly involved) or by changing overall recruit-
a ‘capability’ (to realize a routine) in this paper as ment policies, reward systems, etc., all of which
involves the micro-level.
the ability to repeatedly internalize a pattern of 6. Note that there is a fundamental difference in terms
individual-level external productivity effects. Thus, of levels of analysis between Nelson and Winter
metaphorically the firm possesses a ‘blueprint’ (1982) and later writers in strategic management:
which can carry the firm, without a drop in Nelson and Winter were interested in building
performance, despite turnover (in all the functions (evolutionary) theories that would be rival to the
dominant neoclassical approach with respect to
that are connected to the routine) (cf. Nelson and explaining and predicting outcomes at the level of
Winter, 1982). Testing this idea, while linking it to the industry (i.e. evolutionary price theory) and the
the level of individuals, may involve starting from a level of the economy (i.e. evolutionary growth
certain sampling frame of firms and search for theory). Routines and capabilities were parts of this
stable interaction patterns amongst personnel analytical enterprise, but the aim was not to explain
them per se. This also explains why in Nelson and
(perhaps across functions) through turnover. The Winter’s treatment, quite a lot is packed into the
resulting set of independent variables must then be notion of organization routine, including a variety
related to some measure of sustained performance of behaviours (e.g., heuristics and strategies),
as the dependent variable. Organizations without organizational processes and arrangements, cogni-
routines would need to search and exhibit a tive issues (e.g., ‘organizational memories’), and
incentives (‘truces’). The reason for this all-inclu-
dislocation in interaction patterns. siveness arguably is that ‘routine’ is a catch-all
concept for those collective-level aspects of an
organization that may contribute to the relative
rigidity of firm-level behaviour that is so important
NOTES in evolutionary theory. In contrast, strategic man-
1. Hedström and Swedberg (1996, pp. 296–298) refer agement is mainly interested in explaining and
to arrow 1–3 as ‘situational,’ ‘individual action,’ and predicting competitive advantage, that is, a phe-
‘transformational’ mechanisms, respectively. Hodg- nomenon that is placed on a level of analysis below
son and Knudsen (2004, Section 7) call arrow that of the industry (or the economy), namely the
‘downwards causation.’ level of the firm.
2. Note that this point does not concern whether the 7. In their discussion of routines, Cohen et al. (1996)
explanandum can be placed on the macro level. echo this definition almost verbatim when they
Many (most) explananda in social science are placed define a routine as ‘... an executable capability for
at this level (Coleman, 1990, p. 2)}notably, most of repeated performance in some context that has been
the phenomena that the strategic management field learned by an organization in response to selective
seeks to explain. pressures’ (Cohen et al., 1996, p. 683).
3. Moreover, it can be argued that for pragmatic 8. It has been suggested, however, that there is a
reasons it is often times justified to do research as if hierarchy in firms involving routines, capabilities
arrow 4 causation existed. Thus, Stinchcombe (1991, and dynamic capabilities and that routines, repre-
pp. 379–380) argues that ‘[w]here there is rich senting ‘static’ sequences of actions, are somehow at
information on variations at the collective or the bottom of this hierarchy (e.g., Teece et al., 1997;
structural level, while individual-level reasoning (a) Winter, 2003).
has no substantial independent empirical support and 9. For example, Nelson and Winter (1982, p. 107)
(b) adds no new predictions at the structural level that assume that routines represent organizational truces.
can be independently verified, theorizing at the level 10. It is possible to conceive of units of analysis (e.g.,
of [individual level] mechanisms is a waste of time.’ groups) lying between the firm and the individual–
4. For an elaboration of mechanism-based explanation and, thus there are group externalities– but we
for a management audience, see Felin and Foss abstract from this complexity as group-level phe-
(2006). There is a huge literature in the theory of nomena also invite reduction to the individual level.
science on the nature and role of mechanisms in We use the term ‘productivity effect’ to cover all
explanation. The interested reader may consult possible functions.
Cartwright (1989), Bunge (1997), Glennan (1996), 11. However, some scholars pack much more into these
and Machamer et al. (2000). notions, see, e.g., Levitt and March (1988) for an
5. For example, a correlation between collective extremely expansive definition of routines.
culture and collective outcomes inherently tells the 12. Of course, this might be a surrogate for a manage-
manager very little of what should be done to ment group.
change culture. Similarly, it makes little sense to 13. This assumption is a very strong simplification and
argue that managers can directly intervene on the therefore also a significant limitation of the analysis.
level of, for example, capabilities. Perhaps, however, Essentially we work with a n-person PD game, but
managers can influence capabilities, for example, by solve the game by dictatorial fiat, introducing an

Copyright # 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 489–502 (2008)
DOI: 10.1002/mde

exogenous routine which captures the extent to Abell P. 2003b. The role of rational choice and narrative
which the firm internalizes externalities. In the action theories in sociological theory: the legacy of
present analysis, the routine is not the explicit result Coleman’s foundations. Revue Francaise de Sociologie
of individual action and interaction. A full analysis 44: 255–274.
would also require incorporating the left-hand side Alchian AA, Demsetz H. 1972. Production, information
of the diagram in Figure 2. However, considerations costs, and economic organization. American Economic
of space prevent such an analysis (for an attempt to Review 62: 772–795.
model routine emergence, see Dosi et al., 1999). Argote L, Ingram P. 2000. Knowledge transfer: a basis
14. A pertinent question is why a routine, if it is for competitive advantage in firms. Organizational
relatively easily replicable, is best coordinated in an Behavior and Human Decision Processes 82: 150–169.
organization rather than by markets or by multi- Barney JB. 2001. Is the resource-based view a useful
lateral or distributed bilateral bargaining (Coase, perspective for strategic management research? Yes.
1937). The preliminary answer is that the efficiency Academy of Management Review 26: 41–54.
losses introduced by averaging in organization Becker MC. 2004. Organizational routines: a review of
procedures are less than those associated with these the literature. Industrial and Corporate Change 13:
alternative mechanisms. We leave the exploration of 643–678.
this for treatment elsewhere. Boudon R. 1998. Social mechanisms without black
15. One could go on to study productivity losses by boxes. In Social Mechanisms: An Analytical Approach
introducing averaging under different assumptions to Social Theory, Hedstrom P, Swedberg R (eds).
about the distribution of these effects across the Cambridge University Press: Cambridge, MA;
individuals/positions in the routine. For instance, if 172–203.
they are distributed normally then averaging will Bunge M. 1997. Mechanisms and explanation. Philoso-
not introduce significant distortions; however if they phy of the Social Sciences 27: 410–465.
were to follow a power distribution (which they may Camerer C, Knez M. 1996. Coordination, organiza-
well if the structure of external effects contains hubs) tional boundaries and fads in business practice.
then the average will not capture the impact of the Industrial and Corporate Change 5: 89–112.
effect very well. Cartwright N. 1989. Nature’s Capacities and their
16. Thus, it is assumed that the routine will always Measurement. Oxford University Press: Oxford.
improve output relative to the prisoner’s dilemma Coase RH. 1937. The nature of the firm. Economica 4:
output (i.e. the minimum output), but will never be 386–405.
able to reach the first-best output level (the level that Coff R. 1999. When competitive advantage doesn’t lead
could be reached if information was symmetric). to performance: resource-based theory and stake-
17. Thus, R solves (sub-optimally) a repeated game with holder bargaining power. Organization Science 10:
turnover of actors. Note that individual incentives 119–133.
will not do this in a repeated or one-shot situation. Cohen MD, Burkhart R, Dosi G, Egidi M, Marengo L,
Repeated game equilibria require a stable popula- Warglien M, Winter S. 1996. Routines and other
tion of players for folk theorems to apply. recurrent action patterns of organizations: contem-
18. This is akin to the familiar team production problem porary research issues. Industrial and Corporate
(Alchian and Demsetz, 1972; Holmström, 1982). Change 5: 653–698.
19. This may be taken as an interpretation of Nelson Coleman JS. 1990. Foundations of Social Theory. The
and Winter’s (1982, pp. 107–112) notion of routines- Belknap Press of Harvard University Press:
as-truces. Cambridge, MA/London.
20. We here deviate slightly from the Coleman diagram; Cowan R, Rizzo M. 1996. The genetic-causal tradition
Rj impacts the coefficient relating the micro-level and economic theory. Kyklos 49: 273–317.
variables X and Y rather than the value of X itself, Cyert RM, March J. 1963. A Behavioral Theory of the
which appears to be Coleman’s intention. This is Firm (1993 edn). Oxford University Press: Oxford.
captured in Figure 3 where the arrow 2 in this Dansereau F, Yammarino FJ, Kohles JC. 1999. Multi-
instance is drawn into arrow 3 rather than into X. It ple levels of analysis from a longitudinal perspective:
does seem sensible to allow macro-variables to some implications for theory building. Academy of
modify the impact of micro-motivational variables. Management Review 24: 346–357.
21. Indeed, much work on routines, particularly in Dosi G, Marengo L, Bassanini A, Valente M. 1999.
economics, apply simulation methods (e.g., Norms as emergent properties of adaptive learning:
Marengo, 1996; Hodgson and Knudsen, 2004; the case of economic routines. Journal of Evolutionary
Gavetti, 2005). Economics 9: 5–26.
Dosi G, Nelson RR, Winter SG. 2000. Introduction: the
nature and dynamics of organizational capabilities. In
The Nature and Dynamics of Organizational Capabil-
REFERENCES ities, Dosi G, Nelson RR, Winter SG (eds). Oxford
University Press: Oxford.
Abell P. 2003a. On the Prospects for a unified social Eisenhardt K, Martin J. 2000. Dynamic capabilities:
science: economics and sociology. Socio-Economic what are they? Strategic Management Journal 21:
Review 1: 1–26. 1105–1121.

Copyright # 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 489–502 (2008)
DOI: 10.1002/mde

Elster J. 1989. Nuts and Bolts for the Social Sciences. Reader on Complex Organizations. Amitai Etzioni
Cambridge University Press: Cambridge. (ed.). Holt, Rinehart and Winston: London.
Felin T, Foss NJ. 2006. Individuals and organizations: Leibenstein H. 1987. Inside the Firm: The Inefficiencies of
thoughts on a micro-foundations project for strategic Hierarchy. Harvard University Press: Cambridge,
management and organizational analysis. Research MA.
Methodology in Strategy and Management 3: 253–288. Leijonhufvud A. 1968. On Keynesian Economics and the
Felin T, Hesterly WS. 2006. The knowledge-based view, Economics of Keynes. Oxford University Press:
nested heterogeneity, and new value creation: philo- Oxford.
sophical considerations on the locus of knowledge. Levitt B, March JG. 1988. Organizational learning.
Academy of Management Review 32: 195–218. Annual Review of Sociology 14: 319–340.
Foss NJ. 2003. Bounded rationality and tacit knowledge Lippman SA, Rumelt RP. 2003a. A bargaining perspec-
in the organizational capabilities approach: an evalua- tive on resource advantage. Strategic Management
tion and a stocktaking. Industrial and Corporate Journal 24: 1069–1086.
Change 12: 185–201. Lippman SA, Rumelt RP. 2003b. The payments
Foss K, Foss NJ. 2005. Value and transaction costs: how perspective: micro-foundations of resource analysis.
property rights economics furthers the resource-based Strategic Management Journal 24: 903–927.
view. Strategic Management Journal 26: 541–556. Lucas RE. 1977. Understanding business cycles. Carne-
Gavetti G. 2005. Cognition and hierarchy: rethinking gie-Rochester Series on Public Policy 5: 7–29.
the microfoundations of capabilities’ development. Machamer P, Darden L, Craver CF. 2000. Thinking
Organization Science 16: 599–617. about mechanisms. Philosophy of Science 67: 1–25.
Glennan SS. 1996. Mechanisms and the nature of Makadok R, Barney J. 2001. Strategic factor market
causation. Erkenntnis 44: 49–71. intelligence: an application of information economics
Grossman S, Hart O. 1986. The costs and benefits of to strategy formulation and competitor intelligence.
ownership, Journal of Political Economy 94: 691–719. Management Science 47: 1621–1638.
Hayek FA. 1952. The Counter Revolution of Science. Mäki U. 2001. The way the world works. In The
University of Chicago Press: Chicago. Economic World View: Studies in the Ontology of
Hedström P, Swedberg R. 1996. Social mechanisms. Economics, Mäki U (ed.). Cambridge University
Acta Sociologica 39: 281–308. Press: Cambridge.
Hedstrom P, Swedberg R. 1998. Social mechanisms: an Marengo L. 1996. Structure, competence, and learning
introductory essay. In Social Mechanisms: An in an adaptive model of the firm. In Organization and
Analytical Approach to Social Theory, Hedstrom P, Strategy in the Evolution of the Enterprise, Dosi G,
Swedberg R (eds). Cambridge University Press: Malerba F (eds). Macmillan: London.
Cambridge; 1–31. Miller G. 1992. Managerial Dilemmas. Cambridge
Hodgson G, Knudsen T. 2004. The complex evolution University Press: Cambridge.
of a simple traffic convention: the functions and Nelson RR, Winter S. 1982. An Evolutionary Theory
implications of habit. Journal of Economic Behaviour of Economic Change. Harvard University Press:
and Organization 54: 19–47. Cambridge, MA.
Holmström B. 1982. Moral hazard in teams. Bell Stinchcombe A. 1991. The conditions of fruitfulness of
Journal of Economics 13: 324–340. theorizing about mechanisms in social science. Philo-
Kaplan S, Henderson R. 2005. Inertia and incentives: sophy of Social Science 21: 367–388.
bridging organizational economics and organizational Teece DJ, Pisano GP, Shuen A. 1997. Dynamic
theory. Organization Science 16: 509–521. capabilities and strategic management. Strategic
Kogut B, Zander U. 1992. Knowledge of the firm, Management Journal 18: 509–534.
combinative capabilities, and the replication of Ullmann-Margalit E. 1977. Invisible-hand explanations.
technology. Organization Science 3: 383–397. Synthese 39: 263–291.
Lane D, Lubatkin M. 1998. Relative absorptive capacity Winter SG. 2003. Understanding dynamic capabilities.
and interorganizational learning. Strategic Manage- Strategic Management Journal 24: 991–995.
ment Journal 19: 461–477. Zollo M, Winter SG. 2002. Deliberate learning and the
Lazarsfeld PF, Menzel H. 1970. On the relation between evolution of dynamic capabilities. Organization
individual and collective properties. In A Sociological Science 13: 339–352.

Copyright # 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 489–502 (2008)
DOI: 10.1002/mde

You might also like