You are on page 1of 35

705874

2017
ASRXXX10.1177/0003122417705874American Sociological ReviewMcDonnell

American Sociological Review

Patchwork Leviathan: How 2017, Vol. 82(3) 476­–510


© American Sociological
Association 2017
Pockets of Bureaucratic DOI: 10.1177/0003122417705874
https://doi.org/10.1177/0003122417705874
journals.sagepub.com/home/asr

Governance Flourish within


Institutionally Diverse
Developing States

Erin Metz McDonnella

Abstract
Within seemingly weak states, exceptionally effective subunits lie hidden. These high-
performing niches exhibit organizational characteristics distinct from poor-performing peer
organizations, but also distinct from high-functioning organizations in Western countries. This
article develops the concept of interstitial bureaucracy to explain how and why unusually
high-performing state organizations in developing countries invert canonical features of
Weberian bureaucracy. Interstices are distinct-yet-embedded subsystems characterized by
practices inconsistent with those of the dominant institution. This interstitial position poses
particular challenges and requires unique solutions. Interstices cluster together scarce proto-
bureaucratic resources to cultivate durable distinction from the status quo, while managing
disruptions arising from interdependencies with the wider neopatrimonial field. I propose a
framework for how bureaucratic interstices respond to those challenges, generalizing from
organizational comparisons within the Ghanaian state and abbreviated historical comparison
cases from the nineteenth-century United States, early-twentieth-century China, mid-
twentieth-century Kenya, and early-twenty-first-century Nigeria.

Keywords
development, bureaucracy, political sociology (states), Africa, public administration

Current scholarship on states and develop- decades of wisdom—grounded in what works


ment agrees broadly on two things. First, state within wealthy, industrialized, Western coun-
organizational capacity in low-income coun- tries. Reform efforts sometimes change the
tries is essential to the success of many other letter of the law—a thin external veneer—but
developmental efforts (Acemoglu, Johnson, fail to change the animating spirit of daily
and Robinson 2005; World Bank 2002, 2003;
cf. Zhou, Ai, and Lian 2012). Second, we
have little idea how to engender state organi- a
University of Notre Dame
zational capacity. Efforts to build states’ orga-
Corresponding Author:
nizational capacity are notoriously prone to
Erin McDonnell, Department of Sociology,
failure (Andrews, Pritchett, and Woolcock University of Notre Dame, 810 Flanner Hall,
2013; World Bank 2015). Global experts rec- Notre Dame, IN 46556
ommend organizational reforms based on E-mail: erin.mcdonnell@nd.edu
McDonnell 477

practice (Henisz, Zelner, and Guillén 2005; being located in substantively different minis-
Weber, Davis, and Lounsbury 2009). Yet even terial portfolios, suggesting this is a system-
where organizations earnestly adopt new best atic organizational response to their common
practices, they do not gain the expected ben- position.
efits (Bruhn and Zia 2013; Giné and Mansuri Public sector agencies in contemporary
2014; Karlan and Valdivia 2011). developing countries are typically character-
Scholarship increasingly agrees that fail- ized by low morale, poor productivity, and
ures arise from mismatches between advocated corruption. Civil servants work in decrepit
organizational practices and local institutional conditions, overcrowded into small offices
environments (World Bank 2015). DiMaggio amid towering, unorganized paperwork. Offi-
and Powell (1983:147) famously declared cials are often absent, neglecting government
that “the bureaucratization of the corporation work to attend to informal jobs augmenting
and the state have been achieved.” Largely meager salaries. Others leverage public office
true in the West, this neglects what citizens in for private gain—through bribery, patronage,
developing countries know from experience: or budgetary theft—at the expense of organi-
all contemporary states have formal organiza- zational goals and the public interest.
tions performing administration, but not all This image of dysfunctional public service,
state organizations embody the ethos and although frequently accurate, tells only part of
practices of Weberian-style bureaucracy.1 the story. Monolithically dysfunctional admin-
Instead, many polities—like Ghana—would istrations are the exception, not the rule—
be better characterized as institutional brico- albeit the exception that has long captured
lage, where proto-bureaucratic elements con- popular and academic attention (Evans 1989;
front and contest predominant neopatrimonial Helman and Ratner 1992). Instead, many
practices.2 Paradoxically, I find that highly states regarded as uniformly ineffectual have
effective Ghanaian state organizations eschew great internal variation, with agencies span-
the letter of legitimated global organizational ning a continuum from ineffectual quagmires
practices, and by doing so, they cultivate the to competently achieving organizational man-
underlying bureaucratic spirit that so often dates in the public interest. These state “levia-
eludes reform efforts. This article develops thans” are patch-worked: they are cobbled
the concept of interstitial bureaucracy to together from scarce available resources, with
explain how and why, within such neopatri- organizational diversity sewn loosely together
monial environments, these unexpected and into the semblance of unity. In such states,
unconventional organizational practices adapted Weberian-style bureaucracy exists in
enhance effectiveness.3 interstices—niches within predominantly
Conventional practices advocated in reform neopatrimonial administrations.
efforts are ill-equipped to engender high per- The consequences of variation within state
formance, because they do not address vulner- administrative structures are typically over-
abilities organizations face in a local institutional looked, because a particularly influential
environment lacking a widespread bureau- body of development work is enmeshed in the
cratic ethos. Utilizing a comparison of cases convenient shorthand fiction that “the state”
within the Ghanaian state, I argue that this is is a single, homogenous entity—an assump-
why organizational practices of high-perform- tion woven into cross-national work due to
ing cases differ from ineffectual peer agen- methodologically nationalist limitations of all
cies, but also from Western bureaucracies prominent measures of governance (Kauf-
upon which so much organizational theory is mann, Kraay, and Mastruzzi 2010; Transpar-
based. These cases are not idiosyncratic suc- ency International 2010). Work extolling the
cesses, nor are they dependent solely on developmental virtues of Weberian bureau-
inspired leadership: high-performing niches cracy similarly analyzes the quality and extent
utilize similar organizational practices despite of bureaucracy as a uniform feature of the
478 American Sociological Review 82(3)

state (Evans and Rauch 1999). This notion association with repetitive drudgery. In social
that the state has a monolithic character is contexts where the institutional building
misleading. Scholars’ focus on inequalities of blocks of bureaucracy are comparatively rare,
bureaucratic administration between states I find bureaucratic interstices twist canonical
overlooks tremendous variation within states. features of bureaucracy to mitigate their envi-
The difference in corruption scores between ronmental vulnerabilities and cultivate the
Ghana’s best- and worst-rated state agencies bureaucratic ethos: impersonal administration
approximates the difference between Bel- for the routine, effective, and timely satisfac-
gium (WGI = 1.50) and Mozambique (WGI = tion of organizational duties. Often taken for
–.396), spanning the chasm of so-called granted, routine satisfaction of organizational
developed and developing worlds. duties is an impressive accomplishment of
This spectrum of intra-state administrative socialization, disciplining state agents to miti-
variation goes beyond the poor performance of gate personal interest and coordinate to
corrupt, inept agencies that have captured atten- achieve collective goals.
tion: it also includes agencies performing com-
petently in the public interest despite substantial
obstacles. Existing theories are ill-equipped to Development, States, and
explain the practices of Weberian-style bureau- Organizations
cracy in the context of great administrative Failures of Conventional
variation, because these theories originate from Organizational Reforms
the study of organizational fields where bureau-
cracy is hegemonic. Scholars know volumes In the past ten years, the World Bank (2016)
about bureaucracy as the Goliath of modern spent more than $16 billion trying to reform
statecraft, but what about bureaucracy as David, state organizational capacity, mostly in low-
the underdog and interloper? income countries, where outside experts typi-
The challenges of being embedded in cally recommend changes based on what
adverse environments give rise to what I call works in wealthy, industrialized, Western
interstitial bureaucracy. Building on Mann’s countries—with little success. There are myr-
(1986) work, interstices are social niches iad normative critiques of development practi-
within a larger institutional field, distinct- tioners’ unreflexive export of Western ideals
yet-embedded subsystems characterized by and models as the means and ends of develop-
practices inconsistent with—although not ment, including neoliberalism. Newer critiques
necessarily subversive to—those of the domi- are further grounded in practical concerns:
nant institution. The characteristics of inter- such institutional exports rarely work in the
stitial bureaucracy differ from the hegemonic ways they were intended. Even development
bureaucracy upon which most theory is based, practitioners now recognize “building capa-
because the bureaucratic practice and ethos, bilities of human systems—including that
when locally non-dominant, must confront human system called ‘the state’—has proven
extant institutions and beneficiaries of the much more difficult [than building infrastruc-
status quo, which shapes its character. The ture]” (Andrews et al. 2013:234).
dual cultural competence associated with External and top-down reforms encounter
fruitful innovation (Uzzi and Spiro 2005; limitations, even without translating institu-
Vedres and Stark 2010) is embodied in the tions across borders. External actors may suc-
interstitial structural position of “withinness,” cessfully impose easily observable and
which affords institutional betweenness, ena- enforceable de jure changes, but struggle to
bling creative hybridization. Canonical fea- create bona fide reform that changes de facto
tures of hegemonic bureaucracy include practices (Bromley and Powell 2012; Hallett
hierarchical coordination, discrete jurisdic- and Ventresca 2006; Henisz et al. 2005;
tions, stable promotion from within, and an Weber et al. 2009). This results in a façade
McDonnell 479

familiar in many developing states, layering a including state strength, state capacity, and
veneer of administrative functionality over governance quality. Much cross-national work
inner chaos that haunts reform efforts. Ana- analyzing bureaucratic diffusion, state strength,
lyzing the problematic emergence of modern, capacity, and governance quality reifies the
rationalizing states, Scott (1998:309) docu- state as a singular, homogenous organizational
ments “the natural and social failures of thin, entity, within which the administrative charac-
formulaic simplifications imposed through teristic of interest is treated as uniformly dis-
the agency of state power.” Top-down reforms tributed (Knack and Keefer 1995; La Porta et
replicate abstracted best practices models that al. 1999; Meyer et al. 1997). In this work, all
are intelligible to external actors but fail to prominent cross-national indicators of govern-
address local challenges. Reformers impose ance capacity are measured at the national
their familiar Western bureaucratic models— level (e.g., Kaufmann et al. 2010; Transpar-
which depend on tacit knowledge unfamiliar ency International 2010). In fairness, data limi-
to recipients—while overlooking the unknown tations constrain cross-national statistical
unknowns of local conditions that recipients inquiries to a methodologically nationalist
confront daily. frame. Yet eliding methodological limitation
Organizational reforms predicated on what with empirical reality has consequences: schol-
works elsewhere may fail even when local ars and practitioners have not been sufficiently
actors enthusiastically adopt new practices. concerned with how variation in foundational
Some surprising failures of organizational organizational differences conditions the effec-
reforms are clearest in private enterprises, tiveness—or even appropriateness—of con-
where the outcome of profit provides a crisp ventional Western organizational practices.
evaluation metric. Even well-funded business Organizational practices embodied in reforms
training efforts typically fail at the primary are tacitly predicated on a thoroughly bureauc-
aims of increasing profitability or scale (Blatt- ratized institutional field, which is often lack-
man and Ralston 2015; McKenzie and Wood- ing in low-income states, precisely the states
ruff 2013). Crucially, this is not because local targeted by reform efforts. Thus, an influential
entrepreneurs fail to adopt the organizational body of development scholarship and practice
best practices trumpeted in training, such as has failed to appreciate how variation in organ-
inventory management. Even where trainees izational capacity in the local environment
adopt canonical Western business practices, the may profoundly shape the kinds of organiza-
practices do not improve profitability (Bruhn tional practices that are locally effective.
and Zia 2013; Giné and Mansuri 2014; Karlan
and Valdivia 2011). Randomized controlled tri-
Uncovering Variation within States
als highlight the importance of localized knowl-
edge, which differs substantially from the Political sociologists have long critiqued
organizational content typically covered in some analyses for “exaggerating the unity of
training (Brooks, Donovan, and Johnson 2015). the state and the degree to which state offi-
This suggests there may be a fundamental mis- cials are aware of and serve ‘its’ interests”
fit between organizational practices conveyed (Hooks 1990:363; see also Quadagno 1987).
in training and the local environment within Such reification overlooks complexity in the
which these organizations operate. One of the functioning of states: “viewing the state as a
foremost unappreciated challenges facing single actor risks subsuming sprawling, com-
actors building effective organizations is that plex concatenations of governing institutions
they must cope with disruptions from the sur- under one presumptively unified bureaucratic
rounding high-variation, incompletely bureauc- apparatus” (Morgan and Orloff 2016:15).
ratized organizational field. Newer historical, case-based, and qualitative
State organizational capacity is intricately political sociology acknowledges the state is
interrelated with—and often seen as a founda- not organizationally homogenous (Adams,
tion of—a number of more abstract concepts, Clemens, and Orloff 2005). The powerful
480 American Sociological Review 82(3)

image of a coherent state entity is a social presumed mechanism of buying disciplined


construction accomplished despite “the actual expertise with higher revenues. Gorski’s
practices of its multiple parts” (Migdal (2003) work on the importance of a “discipli-
2001:16). Even scholars canonically associ- nary revolution” in society prefiguring state
ated with state-centered approaches argue that capacity calls into question the ability to will
“there may be insulation or contradictions into being disciplined and effective bureau-
among different kinds of state capacities” cratic agents with only finances.
such that “possibilities for state interventions Alternatively, institutionalist scholars are
of given types cannot be derived from some increasingly analyzing variation among state
overall level of generalized capacity or ‘state agencies and within organizations, focusing
strength’” (Evans, Rueschemeyer, and on how different institutional logics conflict
Skocpol 1985:352–53). Advancing under- within organizational environments (Clemens
standing of institutional change therefore and Cook 1999). Contestation arises when
requires disaggregating the state into com- different agencies—even subunits within
plex, interrelated systems of agents and agen- organizations (Binder 2007; Tilcsik 2010)—
cies (Steinmetz 1999). This scholarship draws have divergent interests due to different pro-
attention to variation within states, but it stops fessional identities (Fligstein 1987; Heimer
short of considering how within-state varia- 1999), organizational goals, and clienteles
tion is consequential for practices that can (Chibber 2002; Gilbert and Howe 1991).
influence the effectiveness of organizations— Conflicting intra-organizational institutional
that is, for precisely the content organiza- logics may arise from cohorts with different
tional reform efforts so frequently target. professional training. For example, in post-
Some political sociology examines variation Communist agencies, a newer cohort of
in organizational capacity, but this work pri- reformists trained in data measurement and
marily views variation as tied to different geo- cost-benefit analysis conflicted with the
graphic locations and scales within national communist-minded old guard over subsidiz-
boundaries, attending to how levels of govern- ing unprofitable organizations (Tilcsik 2010).
ment—federal, state, or municipal—have vary- In Binder’s (2007:568) analysis of conflicting
ing preferences and capacities (Dobbin 1994), institutional logics in public service provi-
including subnational variation (e.g., Singh sion, the counseling department was more
2015). Similarly, scholars of non-Western state- exposed to clients and therefore motivated by
craft use “brown areas” (O’Donnell 2004) and “protecting the children,” whereas accounting
“urban bias” (Bates 1981) to highlight variation was exposed to government auditors and
in the state’s ability to enact and penetrate, but therefore focused on “keeping the money
they see that variation as fundamentally struc- stream flowing.” Baumgartner and Jones’s
tured by geographic and social distances—for (1993:31) foundational work on venue shop-
example, rural peripheries or urban slums are ping4 in Western polities argues that policy
more difficult to govern. success depends on how symbolic framing
Less work focuses on differences among interacts with a venue’s organizational goals:
agencies and ministries comprising the cen- pesticides are viewed differently by an agri-
tral state apparatus. Some prominent histori- cultural committee, which cares about
cal work on early-modern state-building “increasing farmers’ profits,” than by an envi-
tacitly acknowledges the heterogeneity of ronmental group, which focuses on “the nega-
state agencies when emphasizing fiscal tive health effects.” Chibber (2002:957)
organizational capacity as the first-mover and similarly sees the functional division of min-
harbinger of broader state capacities (Levi isterial portfolios driving variations in goals,
1988; Tilly 1990). However, the emphasis on observing that inter-agency conflicts arise
tax authorities leaves unclear how strong even when agencies act in the public interest,
organizational capacity emerges within tax or because “the state is a complex amalgamation
other government portfolios, except for the of agencies, charged with distinct functions.”
McDonnell 481

Such state scholarship primarily views log- cross-national variation in state organiza-
ics as derived from shared professional identity tional capacity for a host of outcomes, from
or organizational goals—counseling depart- conflict to economic growth (Centeno 1997;
ments protect children, agricultural ministries Evans and Rauch 1999; Evrensel 2010; Huber
promote farmer productivity. Conflicts arise and McCarty 2004). Yet this parallel interest
from variation in interests, derived from agen- in technical organizational effects has not
cies’ or units’ official mission. This reflects a enjoyed the same attention to internal varia-
presumption, built into much of Western tion. Scholarship on “developmental states”
organizational theory, that individuals will reignited the sociology of bureaucracy, high-
identify with and align their efforts to their lighting the critical role of bureaucratic
office: if hired as a counselor, they will try to administration in marshaling unprecedented
protect children. This presumes a widespread economic and social development in previ-
bureaucratic ethos as a routine part of social ously poor countries, like South Korea (Evans
habitus. Jepperson and Meyer (1991:208) note 1995; Kohli 2004; Woo-Cumings 1999;
that “myths of the universal collective good” World Bank 1997). Economic development
are essential to generating rationalized organi- depends on organizational features of the
zational structure. Such myths of rationality state administrative apparatus, particularly
and collective good are presumed to be “built whether it possesses characteristics of Webe-
into modern societies and personalities as very rian bureaucracy, including civil service
general values” such that “the building blocks entrance exams, competitive financial com-
for organizations come to be littered around pensation, and long, stable careers within
the societal landscape; it takes only a little civil service (Evans and Rauch 1999).5
entrepreneurial energy to assemble them into a These important contributions to develop-
structure” (Meyer and Rowan 1977:345). ment sociology advanced understanding of the
However, those building blocks are not relationship between development and states
widely available in many non-Western states, as organizational entities. That scholarship
like Ghana. Particularly scarce in these social was primarily built on cases like the “Asian
landscapes is what Weber called the bureau- tigers,” where states used already-effective
cratic ethos: characteristic tendencies to per- bureaucratic administrations to coordinate
ceive the world, think, decide, act, and appear in rapid industrialization. A competent and disci-
particular ways driven by an orientation to the plined bureaucratic administration was a pref-
organization’s goals. The bureaucratic ethos ace to their analytic focus on state-society
embodies the striking transformation of social coordination, a necessary condition that either
attachments that is the backbone of modernity: went largely unexplained or seemed to depend
an affiliation to organization and mission that on the conflux of unique historical conditions.
transcends primordial ties. Following Weber, Comparisons examined catastrophic failures,
the “bureau comprises the social and cultural like the predatory state of the Democratic
conditions of a distinctive and independent Republic of Congo (Evans 1989). Attention to
comportment of the person, one that is basically extremes left unresolved how bureaucratic
non-sectarian in character . . . the office consti- practices function in cases, like Ghana, that
tutes a ‘vocation’ (Beruf )—a focus of ethical are neither extensively bureaucratized nor
commitment and duty, autonomous of and supe- completely failing, which empirically repre-
rior to the holder’s extra-official ties to kith, kin, sent the majority of states globally.
class, or conscience” (du Gay 2005:51). The dominant cross-national paradigm of
corruption research engages the organiza-
tional capacity of states, but again reifies the
Bureaucracy’s Renaissance
state as a unitary entity, using single measures
Sociology of non-Western states and organi- for central state corruption (Evrensel 2010;
zations has considered the consequences of Mauro 1995). Case-based corruption research
482 American Sociological Review 82(3)

acknowledges intra-state variation in organi- patterns suggest it is neither necessary nor


zational capacity, but it tacitly normalizes sufficient: some state functions with high-
bureaucratic administration, presuming cor- level exposure fail to become highly effec-
rupt units are aberrations requiring explana- tive, whereas others with low-level exposure
tion. Corruption research thus overlooks demonstrate uncommonly strong capacities.
cases with high administrative quality, and it The group affiliations approach suggests
generates theoretical mechanisms that cannot an alternative explanation, arguing that state
explain positive deviation within high-variation agencies’ capacity and autonomy depend on
environments. Explanations for corruption whether officials remain strongly identified
center on how certain administrative tasks— with social groups of origin (Carruthers 1994;
especially taxation or customs—allow offi- Migdal 2001). States operate in an arena of
cials lightly monitored access to revenue and competing domestic power sources, like fam-
monopoly control over needed resources, ilies, tribes, and political parties, “in which
which enhances opportunities for corrupt multiple sets of rules struggle for predomi-
practices like bribery (Ferreira, Engelschalk, nance” (Migdal 2001:57). Research on Afri-
and Mayville 2007; Flatters and Macleod can states has long been concerned with how
1995). The near-universal presence of tax and states are affected by the extent to which
customs authorities at the bottom of any list officials remain beholden to ethnic groups
of trusted public organizations across nations (Bates 1974; Bratton and Kimenyi 2008). A
(World Bank 1999–2007) supports the claim concentration of officials from a particular
that certain administrative tasks structurally identity group—if they are strongly identified
enable corruption in a variety of cultural and with that group—may undermine organiza-
institutional contexts. tional capacity by encouraging patronage and
Critically, the converse is not true: the subverting the organizational mission to serve
most effective and trusted ministries vary group goals (Carruthers 1994). This may
widely cross-nationally (Barma, Huybens, explain why some organizations with demo-
and Viñuela 2014; Transparency International graphic concentrations fall prey to patronage,
2010; World Bank 1999–2007). This suggests but not why others succeed. Moreover, there
organizational function is not a sufficient is no clear evidence of a pattern differentiat-
explanation for exceptional performance in ing staff demographics within high-performing
environments where many others fail. Some Ghanaian niches from control cases, nor from
scholars might expect state administrative college graduates generally. Finally, the influ-
functions more connected to the global econ- ence of group affiliations depends critically
omy—such as central banks and finance on the absence of significant esprit de corps,
ministries—to be the most high functioning whereby organizational socialization reori-
(Fourcade-Gourinchas and Babb 2002; ents officials’ identities (Carruthers 1994).
Guthrie 2001). However, empirical patterns Significant esprit de corps is itself an accom-
do not support this intuition: finance minis- plishment of organizational cultural work and
tries are regarded as among the most corrupt therefore endogenous to niche organizational
state agencies in Benin, Ecuador, Madagas- capacity, as I will discuss later.
car, and Zambia, although their central banks Researchers need a new framework for
are well-regarded. The opposite is true in understanding administrative successes. A
Burundi, Guinea, Indonesia, and Slovakia, growing group of scholars are analyzing
where central banks are among the worst- “pockets of effectiveness”—well-performing
reputed public agencies; in Thailand, both organizations in adverse environments
organizations are poorly regarded (World (Barma et al. 2014; Grosh 1991; Leonard
Bank 1999–2007). Exposure to global capi- 2010; Roll 2014; Tendler 1997). They docu-
talism may be one pathway to greater organi- ment high-performing government organiza-
zational effectiveness, however, empirical tions across functions, time periods, and
McDonnell 483

nations, all in environments where most As one of the best-governed states in Africa,
organizations are ineffectual (Grindle 1997). Ghana is just the sort of mid-range case that
These scholars broadly agree that existing has been neglected in conventional research.
theories, focused on explaining public service Ghana possesses a range of state agencies,
failures, are inadequate to explain why some some working effectively in the public’s inter-
public agencies succeed against impressive est, and others approximating the stereotypi-
odds. Grindle’s (1997) comparison study cal image of “third world” civil service
found that neither enhanced compensation mentioned earlier. The coexistence of bureau-
nor organizational function explained differ- cratic and non-bureaucratic administrative
ences between high- and low-performing practices offers insight into the structures and
state agencies. Instead, she found that “man- practices of interstitial bureaucracy. Identify-
agement style” and an “organizational mys- ing the characteristics of Weberian-style
tique” differentiated the strong and weak bureaucratic practices within Ghana thereby
performers, but she concluded it remained yields insight into the cultivation of adminis-
unclear how performance orientation was trative rationality and the nature of state effec-
embedded in organizations. This study, there- tiveness in low-income states more generally.
fore, is a critical extension examining the To select cases of central state organizations
small-group sociological processes and with reputations for effective administration in
organizational features foundational to such the public interest, I interviewed expert observ-
pockets of productivity. ers of the Ghanaian state from 21 organiza-
This article builds on sociological approaches tions, including research institutes, universities,
addressing variation within state administra- international nongovernmental organizations,
tive structures by examining organizational industry groups, and donor organizations.
subunits within the Ghanaian central state. Other selection interviewees included key
Because explanations based on negative varia- actors inside the Ghanaian state: personnel
tion do not explain high-end positive devia- from the Office of the Head of Civil Service,
tions, I examine the range of variation typically the Public Services Commission, the Ministry
overlooked: highly competent groups with of Human Resources, and the Management
strong, positive reputations for effective Services Division, and other high-ranking gov-
administration in the public interest, embedded ernment officials. Selection interviews contin-
in a larger state administrative milieu not con- ued until saturation, when no interviewees
ducive to their functioning. Understanding suggested expert observers or key stakeholder
how trusted and high-functioning groups thrive organizations who had not already been con-
in unsupportive contexts advances scholarship tacted. By cross-referencing public lists of
beyond the presumption that bureaucracy is an organizations consistently invited to stake-
abstract system amenable to superficial diffu- holders’ meetings by the government, I am
sion, instead bringing to light the considerable confident case-selection interviewees repre-
social, cultural, and organizational work that sent the near-universe of significant observers
goes into cultivating and sustaining the bureau- of the Ghanaian government. I interviewed
cratic ethos when it is non-dominant. expert observers about which organizations
were most effective at accomplishing their
mandate and about elements of Weberian
Methods bureaucracy (e.g., meritocratic hiring) least
To disentangle the processes whereby bureau- affected by inimical folk understandings asso-
cracy confronts and contests other institu- ciating bureaucracy with sloth.
tions—core to understanding institutional Based on results from diverse expert
change—sociology needs greater examination observers, I selected four public sector organ-
of states, like Ghana, that are neither bureau- izations with the strongest and most consist-
cratic successes nor over-determined failures. ent reputations for organizational capacity.
484 American Sociological Review 82(3)

Some experts nominated large organizations, I also conducted limited fieldwork with
others specified particular subunits. Within comparison subunits within each larger
nominated large organizations, I refined case organization to better understand the context
selection to core organizational functions and and operations of the focal cases (see Table
selected subunits to study based on agreement 1). These were not pure negative cases—that
between high-ranking internal members and is, they were not intentionally selected for
external observers who mentioned particular absence of the outcome—but rather were
subunits as noteworthy. selected as minimal controls broadly repre-
Selected subunits include (1) the Monetary sentative of the organizations’ operations and
Policy Analysis & Financial Stability Depart- administrative culture. Practical constraints
ment, for short called the Monetary Policy necessitated more limited fieldwork within
Research Department (MPRD), a Bank of controls: I interviewed five or fewer officials
Ghana department that prepared research for per case, and when possible I included obser-
the Monetary Policy Committee, (2) the Pol- vations. To limit extraneous diversity, I
icy Analysis and Research Division (PARD), a selected functionally analogous control cases
group within the Ministry of Finance and when possible.
Economic Planning responsible for generating Following the logic of comparative analy-
and researching fiscal policy initiatives, (3) sis, I focus on characteristics shared across
the Ghana Commercial Courts (GCC), part of diverse primary cases but that are generally
the High Courts—only the Appeals and absent in secondary control cases (for an
Supreme Courts rank higher—dealing exclu- overview, see Table 2). Given space limita-
sively with commercial cases, and (4) Fre- tions, I focus on the case of PARD to illustrate
quency and Spectrum Management (FSM), the main findings and integrate other cases
the field testing and monitoring units of the throughout to demonstrate similarities and
National Communications Authority (NCA), articulate variations. This presentation bal-
which is responsible for licensing, regulation, ances the interests of providing readers
and monitoring of communications. Individ- breadth across cases and depth within a case,
ual interviews covered 60 to 95 percent of while also showcasing the inner workings of
staff in the focal units and lasted at least one the fiscal policy generating unit, a unit of
hour; a number were over an hour and a half. particular substantive interest to political and
With the respondent’s permission, most inter- developmental sociologists.
views were audio-recorded and transcribed. If
respondents declined recording, I took written
notes. Quotation marks indicate recorded ver- Interstitial
batim transcriptions; material enclosed in Bureaucracy In Ghana
apostrophes is from notes taken at the time
(Lofland and Lofland 1995). Participants in “When the [bureaucratic] office is fully
both selected and control cases roughly developed, official activity demands the full
approximate college graduates in Ghana gen- working capacity of the official. . . . For-
erally: most have at least one parent with sec- merly the normal state of affairs was the
ondary school or better education, and they reverse: Official business was discharged as
are from a variety of regions and ethnic groups a secondary activity.” (Weber 1978:958)
(with some under-representation of northern
groups). All participants have at least one col- Observing the ministry area of Accra, Ghana,
lege degree, except support staff personnel in documents a predictable flow of people. Civil
the courts (e.g., transcriptionists) and at NCA servants trickle into offices between 9 and 11
headquarters. Law, engineering, and econom- a.m. Around noon, most exit the ministry
ics degrees predominate in the courts, NCA, buildings for lunch at “chop bars,” eating
and MOFEP/BOG, respectively. steaming bowls of stews with their hands.

Table 1. Overview of Selected and Control Subunits, by Organization

Organization: Ministry of Finance and Econ. High Courts National Communications Bank of Ghana (BOG)
Planning (MOFEP) of Justice Authority (NCA)

Primary Ministry of Finance, responsible Third-ranking Superior Licenses audible, visual, and Central bank of Ghana.
function: for fiscal policy, allocation and Courts in the British- data telecommunication Generates monetary policy,
management of government resources. derived Ghanaian systems; allocates, monitors, issues and regulates currency,
justice system, behind and regulates use of licenses and regulates banking
only the Supreme Court frequencies; monitors quality and nonbanking financial
and Court of Appeals. of services. institutions.

Selected Policy Analysis and Research Division Commercial Courts (GCC) Frequency & Spectrum Monetary Policy Research Dept.
case: (PARD) Management (FSM) (MPRD)
Primary Research and report on economic High court specializing in Monitor and evaluate audible Research and report on
function: impact of fiscal policies proposed commercial law cases. spectrum usage around economic impact of monetary
by Minister of Finance. Some policy the country, including policies proposed by the
generation around fiscal policy issues, field testing with roving Monetary Policy Committee
for example, expediting processing of technical units, reporting on of the Bank of Ghana. Reports
government contract work or budgetary distribution of bandwidth to the Monetary Policy
request approvals. Reports to Minister and service quality. Committee, composed of the
of Finance. Governor of Bank and other
high-ranking govt. officials.

Control Budget Division Regular High Court Admin. Headquarters Banking Supervision Dept.
case:
Primary Budgetary dispersals to ministries for High court hearing all Paperwork administration for Supervision and examination of
function: operational and other expenses, civil and criminal cases licenses and regulations over banking institutions.
including compensation of referred to the high telecommunications, human
government contractors. court, including but not resources management.
limited to commercial
law cases.

485
486
Table 2. Overview of Organizational Characteristics by Ghanaian Cases

Adaptive Missing Managing


Organization Selected and Control Units Autonomy Redundancy Pride Middle Embeddedness
Ministry of Finance and Policy Analysis Research Yes (Achieved) Yes (Ind.) High High Yes
Economic Planning Division
  Budget Division No No Low Low No

High Courts of Justice Commercial Courts Yes (Sponsored) Yes (Unit) High High Yes
  Conventional High Court No No Moderate, Frustrated Low No
Courtroom

National Communications Broadband Monitoring Units Yes (Achieved) Yes (Unit) High Moderate Yes
Authority  Administrative Headquarters No No Low Low No

Bank of Ghana Monetary Policy Research Yes (Formal) Yes (Ind.) High Moderate Yes
Unit
  Banking Supervision Yes (Formal) Yes (Unit) Moderate to High Moderate Yes
Department (BSD)a
a
The Bank of Ghana was the only organization in which internal and external observers nominated the organization in its entirety, as opposed to specific units
within the organization. That is, the Bank of Ghana seems to be a highly effective niche that exists at the scale of the formal organization. As such, the BSD
should not be read as a negative case. Instead, the findings suggest that, as another well-regarded and effective unit, it is noteworthy the extent to which the
BSD has similar characteristics to all the other units, including interstices outside the bank.
McDonnell 487

Some sleep at desks through the heat of early few clusters functioning outside the norms of
afternoon, overshadowed by piles of disor- the larger administrative culture manage to
ganized paperwork. Before formal closing cultivate and sustain these deviant practices.
time, officials begin heading home, or to a How did being embedded within a larger
second job supplementing their income. This administrative field affect the characteristics
is the rhythm of the typical day in the build- of bureaucracy within these subunits?
ings housing the core functions of the Ghana- Although MOFEP’s reputation varied among
ian state. expert observers, PARD was considered, by
At 10 p.m., most office lights have been off both insiders and outside experts, to be a niche
for hours. However, the lights are still ablaze of administrative excellence. My research
in the Policy Analysis and Research Division suggests PARD, like the other bureaucratic
(PARD) of the Ministry of Finance and Eco- niches—but in contrast to the control cases—
nomic Planning (MOFEP). Since before 6 flourished by adapting conventional social
a.m., PARD has been a whirl of action amid technologies of Weberian bureaucracy to
the otherwise-quiet morning. National budget local environmental constraints, enabling
deadlines are approaching, so PARD staff are routine satisfaction of its organizational man-
putting in 14-hour days to ensure the work is date. These adaptations of bureaucratic char-
completed. The division head joins me on the acteristics enabled the niche to carve out a
balcony around the building’s central court- social space of difference, attract and moti-
yard. From our vantage point on one of the vate qualified individuals despite low pay,
highest floors, nearly every office is visible. A and make its work more predictable by
few lights still burn bright, illuminating a defending against disruptions from entangle-
handful of other officials still at desks, typing ment with the larger environment. This data
away at this late hour. section highlights five ways interstitial
‘Still a few lights on,’ I observe. He points bureaucracy inverts classic features of hegem-
to each illuminated room, naming the inhabit- onic bureaucracy.
ants and their department. One is a member of
the Budget Division who used to work for
Carving Space for Distinctiveness:
PARD. In another, the Multi-Donor Budget
Autonomy Instead of Hierarchy
Support team clusters around a desk deep in
conversation despite the late hour. There is a Hierarchy is a canonical characteristic of
pregnant pause, then, still gazing at the few hegemonic bureaucracy; autonomy in part of
lighted offices, he remarks, ‘As for those peo- an organization is considered a source of inef-
ple, they are like us. They come early, they ficiency that reduces coordinated effective-
stay late until the work is done.’ ness and efficiency (Chibber 2002; Weber
This simple proclamation spoke volumes 1978). But when the state-at-large lacks
about the ethos and challenges of interstitial effectiveness and efficiency, autonomy
bureaucracy. In Ghana, staying late is a radical enables the possibility of distinctiveness
badge of the corporatist ethos, only possible where more bureaucratic practices may
when individuals are sufficiently invested in emerge. This section focuses on control over
something beyond the self, particularly when personnel, although as the next section elabo-
lesser levels of devotion are expected. Mem- rates, autonomy may entail various degrees of
bership in this radical subculture was not control over important aspects of organiza-
flaunted. Rather, the camaraderie of difference tional life, including recruitment, resources,
was expressed quietly and inadvertently—the and goals. The Bank of Ghana had formal
lights were on only as a necessary conse- legal autonomy, but the other cases have
quence of working late. Existing scholarship fewer formal opportunities for personnel con-
on African states does not account for such trol: NCA is supervised by the Ministry of
unusual levels of devotion, nor explain how a Communication, GCC falls within the
488 American Sociological Review 82(3)

judicial staffing system, and PARD is subject extensive labor protections, but it is remark-
to centralized civil service hiring through the ably rare and violates hegemonic administra-
Office of the Head of Civil Service (OHCS). tive norms. However, PARD’s director
Because formal autonomy is rare, this sec- reassigned someone to a post outside the divi-
tion analyzes how niches engage cultural and sion for failing to meet the niche’s standards
organizational work to accomplish greater for timeliness, productivity, and work quality.
autonomy over personnel despite low formal This may seem incidental to people with pri-
autonomy. Absent formal autonomy, some vate sector experience in the West, where job
control can be obtained if a “big man” with insecurity is typical, but interviewees under-
status and capital in the larger organizational scored the rarity of transferring a civil servant
milieu takes interest in a niche. For example, for poor performance. Interstitial insiders ani-
the chief justice viewed GCC as a pet project matedly recalled this transfer as an ‘event,’ a
and used his influence to protect hiring cautionary tale in the collective imagination,
appointments, because he viewed the success particularly for new members. Interviewees’
of the GCC as a reflection of his own legacy. retelling of the event underscored it as a
In contrast, PARD achieves quasi-auton- marker that the department was distinctive
omy over personnel through a combination of and “serious.” GCC support staffers were
strategic choices and leveraging political cap- similarly impressed by their own cautionary
ital. PARD leaders used tactics of refusal, tale about a problematic staffer being trans-
poaching, and reassignment to informally ferred for failure to conform to expectations
achieve personnel autonomy. When the list of for punctuality and performance.
recruits from OHCS circulated, other direc-
tors eagerly sought more subordinates, which
Recruitment and Nonfinancial
could increase status in the organization (Par-
Motivations: The Mystery of Missing
kinson 1955). However, PARD preferred to
Middle Managers
take no one if recruits did not meet PARD
standards. PARD would occasionally poach Middle management is the mascot of hege-
officers from elsewhere within MOFEP. monic bureaucracy, born of “office holding as
Recruiting within the ministry provided an a vocation” (Weber 1978:958), entailing a
informal probationary period. Adjoa, a recent secure and predictable “career” with “a sys-
university graduate hired elsewhere in tem of ‘promotion’ according to seniority or to
MOFEP, said, “But when I got there, it was achievement or both” (p. 220). Yet such iconic
like there was nothing there. . . . I come to middle managers are largely absent in Gha-
work, I don’t do anything. So, it’s like I keep na’s interstices. Middle management’s stable
moving on, like going to people asking about bureaucratic career is possible only insomuch
what’s good in other divisions.” She eventu- as that career provides reasonable financial
ally met a senior officer who suggested she security. The harsh fiscal reality for many
was a good fit for PARD. Careful to be politi- low-income countries is that official salaries
cally tactful, people worked behind the scenes cannot compete with business and INGOs
to gain the chief director’s permission to reas- (Rauch and Evans 1997). Therefore, to attract
sign her to PARD. Still others are recruited scarce, high-skilled human resources, niches
after doing national service, affording a year must offer enticements beyond salary.
to acculturate and be evaluated before being Nearly all interstitial insiders reported
officially hired as civil servants. working there despite the salary. The niches
Sanctioning through reassignment also attracted demographically bifurcated clusters:
afforded personnel control. Legal provisions (1) youth drawn to experience and exposure,
and cultural norms make it exceedingly diffi- who could tolerate low pay because of famil-
cult to terminate a civil servant. Reassign- ial support and intended to leverage experi-
ment within civil service does not violate the ence into a more financially secure position in
McDonnell 489

the future; and (2) older, established profes- money but the exposure is huge. And people
sionals from private practice or abroad with that we see right now in very top positions
sufficient financial security, who were drawn around Ghana . . . most have started from
to the prestige of working at the pinnacle of ministries or Ministry of Finance. So, it’s like
their profession, exercising significant influ- a tradeoff now. I’m giving up something,
ence, or making a patriotic gift of their ser- money, a lot of money, to get exposure.” The
vice to the state. By contrast, few mid-life director consciously cultivates beneficial
Ghanaians can afford to make the same trade- exposure, creating great affection and loyalty
offs, because they are expected to provide for among younger staff. Kobena continued:
dependents, including extended kin, which
creates considerable social pressure to avoid “And every meeting that goes on in the
low-paying government work. country that has Ministry of Finance
The modal older member entered public involved, whether a minister’s there or not, my
service later, at higher levels of authority, and colleagues and I can enter the meeting. . . .
aspired to advance within the system to its We go to every place, and that is the expo-
highest levels. For example, one GCC justice sure we have. That is the privilege we have.
noted, “[Where will I be in] five years? And the good thing is our boss not only
Whew. I’m already a high court judge for four wants to have the knowledge, but he diffuses
years now, so probably five years, nine years, it. . . . He pushes, ‘hey, my people, call peo-
I’ll be at the court of appeal.” Because older ple at PARD, tell them to come.’”
members were already near the professional
pinnacle, they saw upper-echelon promotion Young members willingly forgo higher
within public service as offering more of what short-term salaries, seeing the interstice as a
drew them to service in the first place—a long-term investment in experience and expo-
means to achieve even more professional dis- sure, opening up more lucrative careers beyond
tinction and influence—and they foresaw no civil service. When I asked PARD’s youngest
difficulty accomplishing the goal. members where they saw themselves in five
Attracting the brightest young graduates years, only one considered remaining in civil
presents a distinct challenge in a resource- service. Most anticipated leveraging their
scarce environment. Yet, relative to the rest of high-profile experience into graduate educa-
the public service, these niches manage to tion, INGOs, or private sector positions.
attract high-quality university graduates. Kofi Older and younger members shared a
joined PARD fresh from university, emblem- belief that the work itself is valuable—to
atic of the youth perspective: someone beyond themselves—and personally
or professionally interesting. Implemented
“When I joined the civil service . . . someone policy ideas originating from PARD staff
told me, you know how much you’ll be were a source of pride and conversation,
paid? Can you live on 1 million cedis at the including the cell phone tax and a new track-
end of every month? So, I think about it and ing system for paying government contrac-
I was like, oh, 1 million is no money. [pause] tors, which insiders interpreted as valuable
But I also wanted to develop my career. . . . and tangible contributions to the nation.
I wanted that experience exactly about what Whether a communications engineer at the
I had studied in school. So, I said, well, I NCA, a GCC justice, or a newly hired junior
don’t mind. I’m not thinking about money economics officer at PARD, respondents
now. Let’s just get the experience.” described their work as something that gives
them an opportunity to practice their profes-
Younger members were explicitly aware sion and education at the highest levels. Eze-
of the finance-experience tradeoff. Kobena kiel enjoyed the job because of “the quality of
noted, “Well, the ministry is not all about the the work you get.” Daniel echoed that the
490 American Sociological Review 82(3)

primary benefits are “the ability to personally the new commercial courts rather than regular
influence important decisions. And the nature courts. Communication engineers at NCA
of the work itself.” saw themselves as integral to the nation’s
communication system, bridging policy and
practice, and keeping watch for abuses. By
Cultivation and Identification: Pride
contrast, workers in comparison groups rarely
Instead of Cogs in a Machine
mentioned pride in their work, instead citing
All four cases displayed continual boundary the stability of government employment as
work demarcating the interstitial subculture, the primary attraction.
within which different practices and logics
dominate. Talk of pride in comparison rein-
Internal Coping Practices: Adaptive
forced distinctiveness and encouraged posi-
Redundancy Instead of Discrete
tive identification with the niche. Bureaucracy
Jurisdictions
is long-associated with drudgery, alienation,
and diminution of ingenuity. Even Weber Based on established bureaucracy, scholars
([1909] 1956:127), speaking of bureaucratic have long asserted that specialization and
administrators, opined, “it is still more horri- discrete jurisdictions in the division of labor
ble to think that the world could one day be promote orderly and efficient conduct. In this
filled with nothing but those little cogs, little conventional view, each official covers only
men clinging to little jobs and striving towards his own jurisdiction. Workers are not substi-
bigger ones.” tutable. Yet intentional redundancy was pres-
But where the bureaucratic ethos is rare, it ent in all four cases of interstitial bureaucracy,
can be a source of fierce pride. Interstitial and generally absent in the control groups.
niches expressed pride in their unit, the work Redundancy here means work was actively
they did, and how they did it. Rather than managed so each individual or subunit paired
viewing themselves dispiritedly as “cogs,” with another who was nearly identical in
members described PARD as “the engine of skills and project knowledge, and thereby
the ministry,” “the heart tremor,” or “the able to substitute immediately in case of an
backbone of the ministry.” Such expressions absence.6 An external observer might view
frame their work as a dynamic force and an this overlap as fundamentally unbureaucratic.
essential aspect of the work performed in the It defies the taken-for-granted wisdom of
ministry, as expressed by Patience: organizations; no organizational textbook or
management consultant would recommend
“For us, I think, more often, they say this such a strategy (Sorensen 2015). But this is
ministry, the best place to be is PARD. bureaucracy at its finest: a strategic, routin-
Because PARD is the heart tremor. . . . ized response to context, oriented to that core,
PARD is involved in just about everything defining feature of bureaucracy—the routine
that goes on in the ministry. So for us, people satisfaction of organizational goals. Paradoxi-
get to go, we always get to go to meetings, cally, redundancy is how “methodical provi-
we are always involved in a lot of things. We sion is made for the regular and continuous
always get to do a lot of training. Ok, we’re fulfillment of these duties” (Weber 1978:956)
involved in just about everything.” within interstices in the predominant neopat-
rimonial power structure.
Participants in all four cases spontane- Intentional redundancy is an adaptive
ously mentioned explicit pride in their work. social technology enabling productivity and
One GCC justice recalled that since she was a predictability in the context of uncertain-
little girl, she dreamed of serving her country ties—such as disruption of personnel availa-
as a judge. Several GCC support staff like- bility without warning—that come from
wise said they felt honored to be selected for being embedded in larger administrative,
McDonnell 491

social, and epidemiological systems. The duties. So, they, it’s been structured in such
Ghanaian epidemiological system is charac- a way that if I’m not there, someone can
terized by a variety of tropical illnesses that easily just jump in my shoes. And if he is
can strike suddenly, causing prolonged sick not there, I can easily just jump . . . so I
leave and threatening workers’ health and could just go in there and work on proce-
productivity. A flare up of malaria can cause dure as normally as possible.”
an official to unexpectedly take a month off.
The larger sociocultural system within which PARD utilizes a strikingly similar redundancy
interstitial niches operate likewise presents system. Patience described what would happen if
locally distinctive threats to personnel relia- he were unable to report for work at PARD:
bility. Foremost among these are the elabo- Patience: “I have other colleagues, so if I’m not
rate, time-consuming Ghanaian funeral customs. there, they just have to fill in for me, so
For the death of a kinsman, a civil servant they’ll do extra work.”
may be called away on little notice for up to Author: So there are others who know how to
several weeks to celebrate funerary rites. do every element of your job somehow?
Funeral events are held in the family’s village Patience: “Just about, yeah. If, maybe there’s
of origin, potentially requiring a bus ride up about 10. Maybe about—actively involved
to 24 hours away from the capital. One FSM about six [people]. Then there’s another
worker noted, “I have been out of the office person who’s also actively involved in all
for about a month. I lost my mom so I was the six [areas] I do as well.”
out. I just came this week.” By contrast, control cases lacked redundancy.
Another frequent form of personnel disrup- The MOFEP Budget Division, for example,
tion observed during fieldwork came from was organized according to classical division
embeddedness within the larger administrative of jurisdictions, with each desk responsible
culture. In the Ghanaian state, “secondment” is for a non-overlapping portfolio of ministerial
a form of requisitioning staff, temporarily budgets. Although this seems more classi-
transferring a staff member to another unit cally Weberian, it was a poor fit for local
either within or outside the ministry. Second- conditions. When a worker was absent from
ment is a threat arising due to niches’ intersti- his budget desk, work went undone or col-
tial positions. Interstices were targets for leagues attempted to fill in haphazardly with-
secondment from outsiders who needed to out knowing the status of the pending work,
borrow their capabilities, because interstices leading to confusion and delay.
enjoyed quiet reputations for efficiency and
competence—characteristics that were plenti-
Disciplining Outsiders: Managing
ful within the niche but rare elsewhere.
Problematic Embeddedness Instead
Interstices responded to these conditions of
of Hierarchical Order
personnel insecurity by cultivating intensive
jurisdictional redundancy: workers partnered Conventional bureaucracy attends to inter-
with others who replicated their skill sets and office relations primarily as hierarchical sub-
the up-to-date project knowledge required to ordination. However, interstitial bureaucratic
complete the work. They had regular meetings niches are part of the vast state administrative
to keep informed on all aspects of their part- apparatus; their organizational duties often
ner’s work to stay abreast of project knowl- demand contact with outsiders operating
edge. In MPRD, this redundancy was a under different workplace norms and prac-
taken-for-granted aspect of the daily work tices. Therefore, interstitial niches altered
routine. Patrick’s response was typical: their conduct when dealing with outsiders, or
engaged in longer-term projects disciplining
“The way things are organized here in this outsiders who could impinge on the routine
unit, I can do about two different guys’ satisfaction of duties.7 Discipline means not
492 American Sociological Review 82(3)

merely negatively sanctioning, but training apathy, “they just go about it anyhow.”
others into practices beneficial for the suc- Patience repeatedly attempted to acquire data
cessful functioning of the interstitial niche. he was officially entitled to, and he was exas-
Such alteration or disciplining when engaging perated the dilatory outsider would slow the
outsiders were efforts to make their environ- timely completion of PARD’s duties.
ments more predictable—a hallmark of Internal coping mechanisms can mitigate
bureaucratic administration. vulnerability from inter-organizational depend-
Interstitial insiders are cognizant of the encies. Interstitial insiders adjust gathering
cultural work they do to protect their vulnera- techniques, shifting from interstitial ideal prac-
ble position of being embedded in a larger tices to externally valued practices, such as
administrative culture that often undermines appealing to social connections. Junior officers
interstitial practices. Niches are particularly reported relying on elite educational networks
vulnerable to upstream outsider interactions from university. Beyond that, they appealed to
when the niche’s work depends on inputs from “big man” social capital that was respected in
outsiders. Patience gave a common example: the larger administrative environment: they
either asked the subunit’s director to intervene
“There’s a problem when you are going for or emphasized to delaying outsiders that work
data. . . . Maybe they don’t recognize the was required by important figures, like
agency involved and they just go about it MOFEP’s minister. This demonstrates how
anyhow, they go like, ‘come back later for fluidly insiders engage with the larger admin-
it.’ And go back and they still don’t have it. istrative culture through “code switching” ver-
. . . Because when you have a deadline, I naculars, cultural capital, and interaction
have to present something, I need that data strategies (Anderson 2000). Some respondents
to work with.” dropped from the highly erudite English typi-
cally used in the office into thickly accented
The problem is so common that PARD mem- “pidgin” English to replay these exchanges,
bers have a term for it: “chasing.” Another verbally marking the transition from one ethos
PARD member described a similar problem into another.8
with a different data provider: Interstitial insiders occupying positions
accorded power in the larger system—GCC
Kojo: I wanted data from GETFund. I was
sending them a fax; I was chasing on it, so I justices’ legal authority or FSM officers’ reg-
could do some end-of-year poverty expen- ulatory authority—have less oblique means
diture calculations. I was chasing them on it. of disciplining the environment. Both organi-
Author: So how were you chasing it? zations enjoy official sanctioning power, and
Kojo: I called them yesterday. Like the person they are relatively advantaged by how their
is not in. . . . I should call around this time, workflow intersects with interstitial outsiders.
called again [pause] and then they were like NCA and GCC primarily interact with down-
I should call later maybe.” stream outsiders who consume the niche’s
I probed how often routine data required for work outputs. For example, NCA controls
assigned official duties is acquired easily ver- licensing and regulation of audible spectrum
sus chased. Kojo replied, “90 percent I’m use rights, necessary for radio and cellular
chasing.” companies. FSM utilizes mobile technical
PARD’s core mandate is analysis, which is equipment to collect its own data, limiting
critically dependent on externally produced upstream dependencies. If FSM monitoring
data. Patience tacitly distinguished between detects violations, they can issue warnings,
interstitial insiders and outsiders, even seem- impose fines, and, for severe repeated trans-
ing affronted that the outsider did not recog- gressions, threaten licenses.
nize PARD worked differently and therefore Similarly, upon founding, the GCC jus-
handled the request with typical sluggish tices began a long-term coordinated effort to
McDonnell 493

remake their social environment. The justices Temporary reverse cooption occurred
adopted a system of mandatory Alternative through secondment or unofficial voluntary
Dispute Resolution (ADR), where two parties assistance on outside work to ensure it was
meet for mediation before courtroom hear- done promptly and well. Yao recounted seek-
ings. The justices saw this as a core tool for ing data PARD needed and helping an outside
carrying out their organizational mandate— unit manage their data, although this was not
expediently resolving a larger number of part of PARD’s duties: “They conducted a
cases. Statistics from Judicial Service news- survey on, it was on health workers, ok? And
letters document the success of ADR and I helped them with some of the compilation
other GCC organizational innovations: even of some of the data. And I had to print some-
in their first year, GCC handled a larger case- thing out to help them and pull up some data
load with shorter time to resolution than com- at the census’s secretariat.”
parable high courts. Justices report that Reverse cooption can be more significant
lawyers and clients initially resisted ADR and durable. PARD routinely depends on data
because it was unfamiliar and felt less ‘offi- from MOFEP’s Budget Division. Therefore,
cial.’ As GCC quickly earned a reputation for lack of expediency or quality in the Budget
expedient, professional trials and expertise in Division threatens PARD’s productivity.
commercial law, justices leveraged this ben- Eventually, MOFEP’s leadership partitioned
eficial reputation to impose ADR over client the Budget Division. Zakaria, considered
objections. Justices report that most users PARD’s ‘second-in-command,’ was trans-
came to favor ADR after experiencing it. ferred to head one of the budget units, along
GCC also disciplined daily practices affect- with several junior PARD members. Although
ing its goals. I observed judges embarrass the content of budget work was different, the
lawyers who arrived late with incomplete or work style in the budget unit staffed with for-
incorrect paperwork, or who were not properly mer PARD members more closely resembled
attired in black robes and white curled wigs per PARD than that of the other budget unit. Fur-
courtroom protocol. One justice emphasized thermore, PARD and the reverse-coopted
gravely, ‘It won’t happen again, yes?’ If infor- budget unit remained connected through
mal verbal discipline proved insufficient, jus- long-standing interpersonal relationships.
tices could formally sanction lawyers. By The PARD director described how Zakaria
contrast, in the comparison regular high court, remained connected to PARD:
lawyers arrived late, unprepared, or dressed
inappropriately, without sanction. “In fact [he] had been in the policy analysis
Another long-term tactic for managing division all this time until about, about three
inter-organizational dependencies hints at one or four months ago. We still consider him as
possible mechanism for diffusing the practices a member of PARD in the diaspora [chuck-
of interstitial bureaucracy outside of originat- les] because umm, he even seems to be
ing niches. Selznick’s (1949:13) classic study reporting to me still because I had to con-
describes organizational cooption as “the pro- vince him to leave policy analysis and go
cess of absorbing new elements into the lead- because we thought that will help us because
ership or policy determining structure of an he understands the issues. It will help us to
organization as a means of averting threats to coordinate better with the Budget Division.”
its stability or existence.” That is, organiza-
tions respond to threats by incorporating out- This example of reverse cooption suggests
siders into their own leadership structure. By bureaucratic interstices potentially serve as
contrast, interstices may use “reverse coop- sites within which individuals gain lived
tion,” sending insiders out into positions of exposure to the benefits of Weberian-style
leadership in outside units that are otherwise bureaucracy, thereby augmenting the corps of
positioned to disrupt the interstice. bureaucratically experienced and oriented
494 American Sociological Review 82(3)

individuals, that scarcest of all bureaucratic chess when everyone else is playing a differ-
resources. ent game, you have to whittle the pieces your-
self, precious few potential players have ever
seen the game before, and the playing area is
Understanding prone to disruptive bursts of wind.
Interstitial Evans (1995:58) argues that in less-devel-
Bureaucracy oped countries “scarcity, not surfeit, of
bureaucracy underlies ineffectiveness.” Inter-
The organizational practices typifying these stitial bureaucracy extends Evans’s observa-
effective niches are the obverse of classic tion: it is not merely that the human, cognitive,
Western Weberian-style bureaucratic charac- and material resources of bureaucracy are
teristics, because the observed practices are rare. Where they are rare, it matters critically
systematic organizational responses to similar how they are distributed. Clustering a critical
interstitial structural positions. An interstice mass of proto-bureaucratic resources enables
is a distinct niche within a larger field, char- camaraderie born of similarity and synergies
acterized by practices inconsistent with— of concentration that are impossible when
although not necessarily subversive to—the such scarce resources are spread thinly
predominant institutional practices and rules throughout the state. Such clustering also
of the game (Mann 1986). Mann’s (1986) produces distinctiveness vis-à-vis the larger
approach emphasizes structural withinness as environment. These twin needs—clustering
enabling institutional hybridity: social change rare proto-bureaucratic resources and manag-
arises from interactions between old and new. ing precarious distinctiveness—make intersti-
The interstitial lens examines how preexist- tial bureaucracy organizationally distinct
ing, competing practices structuring power from both bureaucratic organizations in the
and behavior affect the formation of intersti- West and poor-performing peer organizations
tial institutions. Bureaucratic interstices fos- in similar environments. The combination of
ter socially innovative or deviant values, within-group similarity and without-group
practices, and cultural tools by clustering a distinctiveness enhances insider identification
critical mass of proto-bureaucratic human, with and commitment to the working group.
cognitive, and material resources. Bureau- This foundation emphasizes organization and
cratic interstices are, effectively, a bureau- mission over primordial commitments, such
cratic subculture—a loosely bounded that work becomes “a focus of ethical com-
numerical minority within a dominant major- mitment and duty” (du Gay 2005:51).
ity, characterized by “a set of modal beliefs,
values, norms, and customs associated with a
But Is It Bureaucracy? From
relatively distinct social subsystem (a set of
Characteristics to Ethos
interpersonal networks and institutions) exist-
ing within a larger social system and culture” Students of classical theory will recall
(Fischer 1975:1323). Even where the sur- Weber’s (1978:1393) formulation of bureau-
rounding institutional milieu is not actively cracy as “characterized by formal employ-
hostile—for example, imposing patronage ment, salary, pension, promotion, specialized
hires on a niche—distinctiveness vis-à-vis the training and functional division of labor,
environment means the environment provides well-defined areas of jurisdiction, documen-
few helpful resources. Contrary to neoinstitu- tary procedures, hierarchical sub- and super-
tionalists’ claim, the sociocognitive “building ordination.” Interstices subvert these conventional
blocks” of bureaucratic organization are not characteristics, but I argue these niches are
“littered around the societal landscape” fundamentally bureaucratic because they are
(Meyer and Rowan 1977:345). Operating as a imbued with the bureaucratic ethos: charac-
bureaucratic interstice is like trying to play teristic tendencies to perceive the world,
McDonnell 495

think, decide, act, and appear in particular important as whether those rules depersonal-
ways driven by an orientation to the organiza- ize administration, lead to more predictable
tion’s goals. Niche members strongly identify outcomes, and encourage decisions based on
with and feel an ethical commitment to the organizational interest rather than personal
office that transcends everyday commitments aggrandizement. Without the bureaucratic
to friends and family. They have distinctive ethos, characteristic bureaucratic features
ways of doing and being in the office—from often fail to achieve the ends of rational, pre-
adaptive redundancy to chasing and staying dictable, effective administration. Conversely,
late—that are oriented toward achieving my data show that in high-variation environ-
rational, predictable, effective, and deperson- ments, characteristics that appear non-
alized administration. Deep reading of bureaucratic actually support the bureaucratic
Weber’s work highlights the centrality of ethos and ultimately achieve rational, pre-
ethos as the core animating force of bureau- dictable, effective, and depersonalized
cracy. The bureaucratic ethos embodies the administration.
distinctively modern transformation of social
attachments, whereby one’s job becomes an
Generalizability
affectively-laden identity category, unreflex-
ively commanding commitment. At the heart In contrast to theory derived from hegemoni-
of bureaucrats’ “distinctive and independent cally bureaucratized environments, analyzing
comportment” is an orientation to the office the bureaucratic ethos as interstitial better cap-
as “a focus of ethical commitment and duty, tures the nature of statecraft where (1) patron-
autonomous of and superior to the holder’s age is the predominant means of organizing
extra-official ties to kith, kin, class, or con- political power, engendering widespread
science” (du Gay 2005:51). assumptions that official positions will be used
Orientation to organizational goals is revo- to enrich oneself and one’s networks, (2) the
lutionary, compared to prior organizational bureaucratic ethos, including other forms of
forms, because it radically depersonalizes social discipline upon which it may be built, is
administration. Instead of personal economic not a routinely available social habitus, (3) but
interest or network ties motivating work in within which exists proto-bureaucratic human,
prebendalism or patrimonialism, bureaucratic cognitive, and material resources that are inter-
officials are oriented to achieving goals on mediate in their intensity, organization, and
behalf of abstract collectives—the organiza- distribution. Such contexts of incomplete insti-
tion, state, or citizens. From paperwork to tutional dominance, where proto-bureaucratic
rules and meritocracy, those characteristics elements are neither too high (hegemonic
are means for cultivating the bureaucratic bureaucracy) nor too low (hegemonic neopat-
ethos and reorienting individual action to col- rimonialism) are conducive to the emergence
lective goals. of bureaucratic interstices.
The oft-cited ideal-typical bureaucratic The mid-range of incomplete institutional
features were social technologies that—under dominance characterizes many formerly colo-
the conditions Weber observed—cultivated nized, low-income states, which underwent
that transformational bureaucratic ethos, facil- decades of external coercion that incom-
itating impersonal administration in the rou- pletely transferred Western institutions,
tine satisfaction of the organizational mandate. thereby weakening traditional institutions
Ultimately, bureaucracy inures not in the char- without fully supplanting them. For example,
acteristic existence of “paperwork,” but in as in Ghana, state administrations in formerly
what paperwork can accomplish: separation of colonized Bolivia, Malawi, and Indonesia are
personal and professional spheres, increased also characterized by a few distinctively
monitoring, and more effective goal achieve- excellent agencies amid many other ineffec-
ment based on analysis of documented pat- tual or corrupt ones (World Bank 1999–2007).
terns. Extensive formal rules are not as Historically, many now-strong states were
496 American Sociological Review 82(3)

Table 3. National Contexts to Which Interstitial Bureaucracy May Generalize

Hegemonic Mid-Range Institutional Hegemonic


Neopatrimonialism Bricolage Bureaucracy
Angola Argentina Canada
Burundi Brazil Germany
DR Congo Chile Japan
Haiti China (Historical) Singapore
Sierra Leone Dominican Republic South Korea
Ghana Sweden
Indonesia Switzerland
Kenya Taiwan
Malaysia United Kingdom
Nigeria
Rwanda
Senegal
Thailand
Vietnam
Philippines
Uganda
USA (Historical)
Uruguay

Table 4. Overview of Abbreviated Comparison Cases

Time Period Country Agency Abbreviation Key Directora


Mid-nineteenth USA Coast Survey Department CSD Alexander Dallas
century Bache
Early-twentieth China Sino-Foreign Salt Salt Inspectorate Sir Richard Dane
century Inspectorate
Mid-twentieth Kenya Kenyan Tea Development KTDA Charles Karanja
century Agency
Early-twenty-first Nigeria National Agency for Food NAFDAC Doris Akunyili
century and Drug Administration
and Control
a
Refers to the director at the time when scholars indicate that the agency emerged as distinctively
effective; in all cases, the specific organization—or an organization that performed the same state
function—existed prior to becoming highly effective.

once characterized by a critical mass of proto- public interest despite operating in adverse
bureaucratic elements in the context of gen- environments in which neopatrimonialism,
eral prevalence of patronage, including patronage, and poor administration predomi-
Republican-era China and the Jacksonian-era nate (see Table 4). In the United States, during
United States. Table 3 gives a non-exhaustive the Jacksonian-era synonymous with patron-
list of country-cases meeting boundary condi- age, the Coast Survey Department (CSD)
tions for interstitial bureaucracy, based on became “a world-renown scientific agency,
secondary literature and discussion with doubtless producing the best known of all the
scholars with lived expertise in-country. scientific work sustained by the government”
Within these mid-range contexts, as with (White 1954:490). In Republican-era China,
Ghana, a few state agencies or departments the Salt Inspectorate inherited an ad hoc sys-
are distinctively high-functioning, known for tem prone to local capture and “tax leakage,”
competent, impersonal administration in the and yet increased net revenues by 747 percent
McDonnell 497

Figure 1. Visual Overview of the Mechanics of Interstitial Bureaucracy

in its first 10 years, although it “operated in an Autonomy


extremely turbulent, often hostile, environ-
ment,” as much of post-dynastic China The clustering and contradistinction from the
descended into “warlordism” (Strauss 1998: larger neopatrimonial institutional field that
59). In post-independence Kenya, amid infa- characterize interstitial bureaucracy are pos-
mous ethnicized neopatrimonialism, the sible only with sufficient operational inde-
Kenya Tea Development Agency (KTDA) pendence—especially over personnel, tasks,
increased the quantity of production 50-fold and rewards (including status rewards). Con-
and improved quality to command above trary to the erroneous presumption that orga-
world-market prices, achieving high rates of nizations everywhere are already thoroughly
return on investments (Grosh 1991; Leonard bureaucratized, in many non-Western con-
1991). In contemporary Nigeria, amid turbu- texts the bureaucratic ethos is the deviant
lent politics rife with corruption—and while institution, requiring autonomy to insulate it
facing targeted vandalism, arson, and kidnap- from the contrary practices of a patronage-
ping attacks—the National Agency for Food based administrative milieu.
and Drug Administration and Control (NAF- Unusually high operational autonomy was
DAC) halved the rate of counterfeit drugs in present in every observed case of interstitial
circulation, emerging as the agency with by bureaucracy, from PARD’s informally
far the strongest public reputation for effec- achieved autonomy to formal autonomy in
tiveness (Pogoson and Roll 2014). Ghana’s BOG, China’s Salt Inspectorate, and
The remainder of this section will argue Nigeria’s NAFDAC.9 Both formal and infor-
that across considerable variation in time mal autonomy depend on elite support or inat-
period, geography, and state function, these tention, either of which can protect from the
unexpectedly effective organizations possess more sundry effects of political elite interfer-
similar interstitial mechanics (see Figure 1): ence—like patronage staffing. Neopatrimonial
sufficient operational autonomy to control political elites need patronage dumps—organi-
personnel and rewards, enabling clustering of zational dumping grounds saturated with
otherwise rare proto-bureaucratic resources to patronage employment—to maintain power
produce a subculture of bureaucracy that is (Swidler 2016). Large ministries with plentiful
distinctive from the surrounding institutional job opportunities—for example, education or
milieu. That distinctiveness simultaneously public works—afford more patronage opportu-
promotes within-group identification and nities than do smaller, technical units (Schrank
necessitates organizational responses to miti- 2016), like Ghana’s FSM. Opportunities for
gate environmental vulnerabilities. non-legal income, like in customs or traffic
498 American Sociological Review 82(3)

police, can make patronage posts even more bureaucratized domestic private sector, or
attractive. Elites may ignore smaller or techni- working or training abroad, lived experience
cal pockets of central state administration, formulates within such technocrats a largely
which are structurally undesirable for patron- unconscious schema. This schema, or network
age and corruption, thereby affording greater of neural associations, influences how people
autonomy. Alternatively, political elites may perceive and evaluate a situation, making
actively support agencies when elites’ interest some practices and actions more readily think-
in organizational accomplishments trumps able while discouraging others (Lizardo 2017;
competing political or patronage goals, as was Vaisey 2009). Directors with both techno-
the case in KTDA (Leonard 1991). Elites may cratic merit and elite connections possess
voluntarily foster organizational achievement enough trust to gain discretion, enough famili-
if cognitive and cultural understandings frame arity with alternative cognitive schema to pur-
particular state functions as essential to elites’ sue organization achievement over patronage,
symbolically laden nationalist or modernizing and enough leverage vis-à-vis elites to push
projects (Barma et al. 2014). International unconventional organizational practices. For
creditors may also coerce elite interest in example, Alexander Bache—head of CSD—
agency success—for example, credit crises was one of the foremost scientific minds of his
facilitated elite support for tax collection by era, but also from a politically influential fam-
the Salt Inspectorate (Strauss 1998). However, ily that included Benjamin Franklin (White
international pressure alone is often unsuccess- 1954). Charles Karanja, KTDA director, was a
ful (Johnson 2015). Political elites’ (in)atten- credentialed engineer when university gradu-
tion is inflected through prevailing local ates were sparse (Kenya 1960), but also a
interpretations and interests, further explain- personal associate of President Kenyatta.
ing the lack of clear cross-national patterns in Thus, the patrimonial milieu encompassing
effectiveness by state function. interstices may alter even meritocracy—a
Elites may tolerate greater organizational keystone bureaucratic characteristic—toward
discretion when they have interpersonal ties mixed meritocracy, whereby leaders have
with organizational leaders. Interpersonal ties both personal connections to political elites
are substitute disciplinary mechanisms where and a sufficiently rare claim to expertise.
the bureaucratic ethos of discipline on behalf
of an office or abstracted collective good is not
Clustering
a widespread part of the social habitus (Adams
2005). Therefore, even when elites seek tech- Even autonomous state agencies cannot con-
nically competent candidates, prudence sug- trol the skills and tacit socio-organizational
gests presuming recruits hold external loyalties schema of the labor pool (see Figure 2).
in competition with the organization’s best Potential recruits variously possess both
interests. Thus, significant discretion is only subject-relevant technical skills—explicit
tolerated where loyalty is assured. knowledges like econometrics or Excel—and
Simultaneously, relatively rare technocratic often-tacit cognitive schema—socio-organiza-
experience or credentials help orient opera- tional orientations, experiences, motivations,
tional discretion toward organizational achieve- judgments, or practices—with elective affini-
ment. Technocrats possess authority derived ties for the bureaucratic ethos. Within a hege-
not only from elite connections. Threat of monic bureaucratic milieu, these building
dismissal by neopatrimonial elites means less, blocks of organization are (as neoinstitutional-
thanks to technocrats’ relatively plentiful ists presume) helpfully strewn throughout
alternative job opportunities. Ideas about what society; bringing them together takes little
constitutes good work are not wholly derived entrepreneurial effort. Conversely, the few
from patron-client relationships. Whether such individuals who enter civil service within
grounded in domestic professionalized hegemonically neopatrimonial environments
training, experience in a sufficiently exhaust themselves standing alone against the
McDonnell 499

Figure 2. Proto-Bureaucratic Concentration in Labor Pool and Clustering within State


Administration

strong current of neopatrimonialism. Within and further organizing of the bureaucratic


societies in the institutionally bricolaged mid- schema embodied within existing members
range, a modest number of potential recruits through lived experience (see Figure 1).
possess technical skills and proto-bureaucratic Niche practices explicitly orient toward
cognitive elements. But if thinly dispersed, achieving organizational goals, in contrast to
these recruits face the same overwhelming neopatrimonial practices prevalent in the sur-
challenges that frustrate individual effort. rounding milieu. At each stage, these distinc-
Clustering a critical mass of these scarce tive practices simultaneously have direct
resources enables new organizational possi- technical benefits advancing the organiza-
bilities, fostering in-group identification, dedi- tion’s goals and no-less-important indirect
cation, and enhanced effort, culminating in a effects of reaffirming social boundaries
cognitive subculture of the bureaucratic ethos. between niche and milieu. These symbolic
Clustering is the process of (1) recruitment boundaries unsettle within the group the
and pruning, the ongoing selection about otherwise-settled cognitive schema of neopat-
which persons are in (or out) of the group, rimonialism that are so prevalent and practi-
and (2) cultivation, the enriching, thickening, cal outside the group (see Swidler 1986).
500 American Sociological Review 82(3)

Recruitment and pruning. Niches seek transcends face-to-face contacts. Profession-


both job-relevant substantive skills (traditional alization also promotes seeking nonfinancial,
meritocracy) and difficult-to-discern proto- professional status rewards. Individuals
bureaucratic socio-organizational orientations. thereby acquire “orientations which transcend
Unfortunately, these characteristics—like individual rational-instrumental behavior,” a
being “incorruptible” or “hard-working”—are necessary foundation for institution-building
rare and difficult to evaluate except by per- (Rueschemeyer 1986:59), by participating in
sonal knowledge or observation, creating high a sufficiently professionalized occupation.
search costs. This may explain why prior Economics, engineering, and law are all con-
knowledge of candidates is a comparatively sidered highly professionalized fields within
common feature of recruitment within bureau- Ghana, endowing the Ghanaian niches with
cratic interstices. PARD poached observable an advantageous baseline pool.10 However,
hard-workers from within the ministry or the extent of professionalization varies cross-
hired national service interns after a year’s nationally and by occupational field, further
observation. GCC justices were hand-picked explaining why there is not more cross-
by the chief justice for excellence in legal national consistency in the location of inter-
practice, but also for a reputation for a strong stitial bureaucracy by state function.
work ethic and what one interviewee called Niches offer personally costly, high-time,
‘incorruptible character.’ FSM and MPRD high-effort work, such as travel to undesira-
recruitment relied heavily on recommenda- ble, remote locations: for example, FSM
tions from professors at elite training pro- field-monitoring of audible spectrum usage,
grams. At NAFDAC, Akunyili’s hand-picked inspectorate officials collecting salt tax from
management team balanced technical merit rural regions, or NAFDAC inspectors field-
and “proven integrity and commitment to the testing for counterfeit drugs. Those costs are
NAFDAC mission” (Pogoson and Roll balanced principally by benefits that do not
2014:103). Recruitment into KTDA balanced attract all applicants equally: engaging in
“an element of personal loyalty mixed in with high-impact work and (typically nonpecuni-
a preponderance of objective merit” (Leonard ary) status rewards. Consider PARD’s recent
1991:134). Ironically, the seemingly non- graduates seeking exposure to high-profile
bureaucratic practice of recruiting among professional work or GCC justices drawn to
known associates may be an essential aspect practicing at the pinnacle of their profession.
of interstitial bureaucracy, because direct The cumulative structure of costs and benefits
observation or interpersonal knowledge may tends to disproportionately attract profession-
be the only effective way to manage search oriented candidates and selects out applicants
costs and discern rare and difficult-to-measure merely seeking stable government jobs,
characteristics. which can be less effortfully had elsewhere.
Structural challenges from the larger institu-
Nonpecuniary motivations. Where tional milieu may shape the pool of applicants
rationalizing myths are not part of the wide- for whom the professional enticements exceed
spread social habitus, state-builders instead the time and financial opportunity costs—for
appropriate other forms of social disciplining example, kinship obligations placing a finan-
that habituate individuals to working toward cial burden that overwhelms even profession-
collective goals (Adams 2005; Gorski oriented midlife Ghanaians.
2003)—in contemporary poor states, the fore- The concentration of rare proto-bureau-
most example is sufficiently professionalized cratic socio-organizational orientations was
occupations (Abbott 1988; Leonard 1991). further defended by pruning niche insiders
Professionalization accustoms members to who failed to conform to the distinctive com-
identifying with and orienting toward advanc- portment of the niche. Pruning improved pro-
ing the status of an abstract collective that ductivity by directly eliminating unproductive
McDonnell 501

insiders, but—as in PARD and GCC—also Insiders experience their niche as distinc-
further distinguished niches from business- tive vis-à-vis the larger environment—in par-
as-usual in the surrounding milieu, where ticular, the niche offers a positive source of
punishment for poor productivity was identity that sharply contrasts the spoiled
unthinkable. Dismissals for underperformance identity of public servants broadly—which
or corruption were similarly noteworthy cultivates insider loyalty and shores up organi-
within NAFDAC—where Akunyili cleaned zational boundaries against neopatrimonial
house after her appointment—and KTDA, logics. Individuals discover shared common-
where dismissal explicitly flaunted ethni- alities through interaction and strengthen
cized patronage politics as Karanja “applied those perceived commonalities within groups
the principle as rigorously to his relatives (Wohl 2015). Through ongoing interaction,
and the politically connected” (Leonard including social disciplining (especially of
1991:133). new members), perceived commonalities
reach a relatively stable equilibrium. Group
Cultivation. Lived experience within the judgment, taste, style, and appropriate or val-
niche amplifies proto-bureaucratic common- ued practices are reinforced in regular interac-
alities, aggregating into increasingly ordered tion within a bounded social space, comprising
schema as group participation routinely reaf- a group style, or “recurrent patterns of interac-
firms a proscriptive and prescriptive bureau- tion that arise from a group’s shared assump-
cratic logic—oriented toward achieving the tions about what constitutes good or adequate
organizational mandate—in contrast to participation in the group setting” (Eliasoph
neopatrimonial logics. Newcomers to PARD and Lichterman 2003:737). Lived experience
recalled with fascination witnessing the direc- within a niche comes to comprise a largely
tor deftly sidestep neopatrimonial influence unconscious schema keyed to the office, a
while avoiding confrontation, and they network of neural associations influencing
marked the encounter as a template for their perception, evaluation, and action (Lizardo
own future action. 2017; Lizardo and Strand 2010; Vaisey 2009).
Each historical case of distinctively high- That which interstitial insiders have in com-
functioning niches involves particular innova- mon, and which makes them maximally dis-
tive technologies, to which scholars attribute tinct from outsiders,11 is the style of
the distinctive success of the agency. GCC comportment oriented toward accomplishing
innovated Alternative Dispute Resolution to the organization’s goals while eschewing pat-
streamline time to adjudication. The Salt rimonial influences. Thus, the distinctive
Inspectorate innovated new salt evaporation comportment that comes to characterize the
techniques to improve quality. The CSD inno- interstice is the bureaucratic ethos: character-
vated mathematical techniques for adjusting istic tendencies to perceive the world, think,
triangulation networks in surveying. KTDA decide, act, and appear in ways driven by an
innovated farm-to-processing techniques for orientation to the organization’s goals.
tea cultivation. The particular content of tech- Paradoxically, this work suggests the
nical solutions varied, but what transcends administration of the office may be carried
cases is that procedures were selected on the out according to “general rules, which are
basis of organizational goal-oriented logics, more or less stable, more or less exhaustive,
and were thereby notably distinct from pre- and which can be learned” (Weber 1978:958),
vailing logics of personal enrichment, political but the bureaucratic ethos upon which that
preference for particular groups, or nurturing administration depends cannot. Just as every
clientelist relationships. Beyond direct techni- contingency of a business relationship cannot
cal benefits, utilizing distinctive goal-oriented be enumerated in a written contract, such that
technologies affirmed the boundary between contracts depend on unspoken, non-contractual
the interstice and the environment. elements (Macaulay 1963), bureaucracy
502 American Sociological Review 82(3)

includes rules and features of administration, challenges to personnel availability—like


but also myriad practices that together com- malaria—are perceived to be too big, dis-
prise the bureaucratic ethos. One reason top- persed, and technically beyond the capabili-
down reforms fail is that the bureaucratic ties of PARD to change, and therefore the
ethos cannot be effectively and exhaustively niche modifies its internal working processes
enumerated such that it can be transmitted by to mitigate the risk, adopting adaptive redun-
mandate. Rather, it must be lived to be learned. dancy. Conversely, particular problematic
behavioral patterns among down-stream
recipients of a niche’s work products—like
Managing Embedded Distinctiveness
punctuality among lawyers at GCC—are
Clustered distinctiveness fosters insider iden- more likely targets of external disciplining
tification and dedication, but it also poses projects that remake the social practices of
unique challenges. Ensuring routine comple- outsiders interacting with the niche.
tion of organizational duties in an environ-
ment that ranges from not conducive (e.g., Internal organizational responses.
lack of time-disciplining habitus means The specific content of internal organiza-
PARD has to chase data) to actively disrup- tional responses is perhaps as varied as the
tive (e.g., personnel secondment or funerary diverse local challenges interstices confront,
travel demands) means niches spend greater yet they all share the core characteristic of
organizational effort mitigating basic vulner- targeting particular local conditions and chal-
abilities, compared to Western organizations lenges with solutions that were relatively
or poor-performing local peers. Bureaucratic uncommon in the local environment (further
interstices within neopatrimonial state admin- underscoring distinctiveness), in contrast to
istrations are embedded in a preexisting sys- one-size-fits-all global best practices. When
tem that is largely beyond their control. State most government officials enjoyed lucrative
interstices are tied to the state administrative patron-client relationships with local elites,
apparatus: state organization precludes free- the Salt Inspectorate instead frequently
market competition, so interstices cannot rotated official postings “to forestall incipient
select alternatives when partners are under- tendencies toward ‘capture’ by local elites”
performing. In Ghana—as in Nigeria, Repub- (Strauss 1998:72). The KTDA utilized prac-
lican China, or Jacksonian America—in the tices that seem the opposite on paper, but
surrounding administrative milieu, sloth, which were similar in responsiveness to their
ineptitude, and patronage are the expected particular set of local environmental chal-
rules of the game that structure initial interac- lenges. Within the intense racially and ethni-
tions between state functionaries and outsid- cally charged context of post-Independence
ers. Interstices must cope with interruptions Kenya, when Kikuyu favoritism was expected
arising from how those expectations structure but resented, KTDA engaged a variety of
how outsiders approach the niche. managers to work coethnically with tea farm-
Niches deal with environmental challenges ers of the same ethnicity. This engendered
by modifying internal practices and engaging trust and made visible the demographically
in long-term projects to remake their physical proportional hiring of management.
and social surroundings. Whether and to what
extent interstices mitigate vulnerabilities External projects of reshaping. Inter-
through internal coping mechanisms or stices may also engage in long-term projects
engage in external disciplining projects to reshape their environment. Such external
depends on structural organizational vulnera- reshaping to serve organizational goals could
bilities (e.g., upstream versus downstream), include infrastructural projects. For example,
as well as the nature, magnitude, and dispersal the Salt Inspectorate remade the relevant
of the challenge. For example, epidemiological physical environment to mitigate external
McDonnell 503

threats to their ability to achieve their organi- organizational sociology of political bodies,
zational mandate: they constructed salt depots particularly in many developing countries.
to frustrate smugglers, and they improved Broadly, it suggests scholars should recon-
road access to facilitate compliance. But most ceptualize states as patch-worked composites
common among all cases are projects of dis- of fractious subunits that may differ vastly in
ciplining routinely encountered outsiders into foundational organizational capacities and the
new ways of interacting with the niche. GCC institutional rules of the game guiding their
justices disciplined lawyers. KTDA disci- actions.
plined tea farmers. The CSD disciplined Interstitial bureaucracy unpacks the micro-
coastal tide observers. foundations of a canonical institution—
A niche’s interstitial structural position bureaucracy—calling attention to the
makes challenging long-term projects to cognitive and cultural elements of institutions
reshape the environment plausible. Signals of that are embodied in what Weber called the
distinctiveness demarcating a cluster—both bureaucratic ethos. It thus follows in the foot-
behavioral and physical, like GCC’s distinc- steps of scholars like Evans (1995), who
tive building—cue outsiders when alternative points to esprit de corps, and Rueschemeyer
schema apply. Clustered distinctiveness (1986, 2005) who calls attention to the non-
affords economies of concentration that pro- bureaucratic elements of bureaucracy. It
vide incentives for learning new practices: draws on advances in cognitive and cultural
whereas lawyers may forgo the effort of sociologies—strangely absent in most macro-
learning new practices to satisfy one judge, a political sociology—to identify cognitively
commercial lawyer cannot readily ignore plausible foundations for how understandings
concerted signals from all commercial law and practices emerge, are shared, and stabi-
specialty courtrooms, which he will revisit lize within groups. This is consonant with
regularly in the future. Moreover, niche insid- institutional and organizational research seek-
ers share the burden of disciplining outsiders ing to “inhabit” institutions (Hallett and Ven-
and support each other through the sometimes tresca 2006), situating institutions within
exhausting work of doing so. The frustration human actors engaged in specific contexts.
and futility of lone effort is illustrated by the Institutions like bureaucracy are not merely
justice in the control courtroom. He espoused material manifestations, such as written rules
identical desires for judicial practice, yet he and regulations, but also a set of orientations,
was routinely frustrated in his efforts. It is not practices, and inclinations grounded in the
merely that the sea of environmental threats lived experience of people. Some scholarship
overwhelms, for interstitial niches face the has examined the global diffusion of the out-
same waves. A clustered, critical mass of dis- ward material manifestations of bureaucracy,
tinctive members with strong internal identi- but I join others who argue we have been
fication can engage processes of external overconfident in the ability of these materials
change that would otherwise overwhelm indi- to cultivate the animating bureaucratic ethos.
vidual efforts. This distinction matters because without
the bureaucratic ethos, classic bureaucratic
features often fail to achieve the ends of
Conclusions rational, effective administration—an under-
Patchwork Leviathan advances a research appreciated reason why organizational reform
tradition of state-building largely based on efforts fail. Conversely, in environments
premodern European states (Adams 2005; where the building blocks of rational organi-
Ertman 1997; Tilly 1990), calling theoretical zation are not widespread within social habi-
and empirical attention to within-state varia- tus, characteristics that appear to defy Western
tion in capacity. Within-state variation is a organizational wisdom—like redundancy—
significant empirical reality conditioning the may actually support the bureaucratic ethos
504 American Sociological Review 82(3)

and ultimately achieve rational, predictable, distinctive work group—particularly along


effective, and depersonalized administration. vectors that maximize out-group distinctions.
Adopting unconventional organizational Over time, this culminates in a subculture of
practices is paradoxical from the perspective bureaucracy that is distinctive from the sur-
of existing scholarship, which expects organi- rounding neopatrimonial environment. That
zations to adopt globally legitimated models, distinctiveness simultaneously promotes
but also anticipates decoupling and thus an within-group identification and calls for
eventual failure to reproduce the underlying active efforts to mitigate environmental vul-
spirit of reforms. By contrast, bureaucratic nerabilities. These conditions are both struc-
interstices adopt practices that do not con- tural and interactional. Recruitment through
form to global organizational wisdom, but in networks is interpersonal, but it is also
so doing, they cultivate the underlying spirit affected by structural factors shaping the
that often eludes reform efforts. availability and distribution of proto-bureau-
This article explains both how interstitial cratic tendencies within the population. A
bureaucracy differs from hegemonic bureau- seemingly structural condition, like a position
cracy and why. Hegemonic bureaucracy may of autonomy or distinction, is ultimately the
be characterized by the hierarchical coordina- culmination of concatenations of interper-
tion of discrete jurisdictions, staffed by offi- sonal interactions unfolding over time, even-
cials motivated by stable and financially tually resulting in a set of stabilized practices
secure employment throughout a lifelong in which those inside and outside the niche
career steeped in drudgery. By contrast, the feel and act as though the niche is autono-
Ghanaian interstices were defined by high mous and distinctive.
operational autonomy, organized by adaptive Interstitial bureaucracy raises compelling
redundancy, spent inordinate organizational new research directions, especially for schol-
effort managing external environmental dis- ars concerned with development, states,
ruptions, and staffed by officials with nonpe- bureaucracy, and organizations. Investigating
cuniary motivations who did not anticipate or an uncommon phenomena like bureaucratic
even desire stable lifelong employment, but interstices faces methodological challenges,12
were bursting with pride. Interstitial bureau- so theory-generating research is an important
cratic organizational characteristics differ first step to guide future research. Why does
from canonical Weberian bureaucracy, because interstitial bureaucracy develop in some parts
the adapted characteristics are systematic of the state but not others? Under what condi-
responses to the interstitial position. This tions do these niches endure or become over-
position demands novel organizational prac- whelmed by the larger environment? What
tices to cluster and cultivate resources that are might enable their distinctive ethos and prac-
not widely available, and to manage disrup- tices to spread within the state? Under what
tions arising from interdependencies with the conditions does interstitial bureaucracy
adverse surrounding milieu. emerge as organization-wide (e.g., NAF-
Sufficient operational autonomy matters DAC) or as a subunit (e.g., PARD), and how
because it conveys control over personnel and does the scale affect particular dynamics?
rewards, affording deviation from the status Niches of corruption may also be based on
quo in the surrounding neopatrimonial envi- autonomy enabling selection and cultivation
ronment. Autonomy enables clustering of that yields a concentration of distinctive per-
otherwise rare proto-bureaucratic resources, sons, typified by strong insider identification
through intentionally recruiting for proto- (although based perhaps on shared risk) and
bureaucratic orientations and recruit-side the careful management of external vulnera-
selection effects. Clustering is further culti- bilities. How might the interstitial process
vated as niche insiders come to identify and vary when explaining pockets of normatively
conform with the standards that typify their “negative” deviations—like a particularly
McDonnell 505

corrupt police department—embedded within protect them from political predation (Car-
largely bureaucratic environments? There is penter 2001), helping them endure.
some evidence that clustering distinctive per- Conditional on niche survival, diffusion of
sons may be correlated with particularly the distinctive ethos of interstitial bureau-
effective niches in early-modern European cracy may be most likely to occur through
states (Aylmer 1961), but how does the world- inter-agency transfer of experienced intersti-
historical context—lacking globally prevalent tial insiders, because the bureaucratic ethos is
bureaucracy—affect the process? grounded in lived experience. This suggests
This article explains how bureaucratic diffusion may be more likely to other units
niches work, by extension suggesting that the whose qualification requirements are similar
locations where bureaucratic interstices will enough to enable personnel transfer. How-
exist are those where the conditions are met.13 ever, transferring a sufficient concentration of
It is a fascinating question, although beyond experienced interstitial insiders to inculcate a
the present scope, to explain how some of new interstice may depend on active coopera-
those conditions themselves came into being. tion from political elites. Where niches sur-
However, several hypotheses emerge for vive on benign neglect from political elites,
scholars interested in identifying patterns in they may arrive at durable interstitiality.
precipitating conditions. Given the impor- Niches that are too identified with particular
tance of clustering a critical mass of numeri- political patrons may become the sacrificial
cally rare distinctive people, hypothetically, lamb of the next administration, leading to
interstitial niches should be more likely in their demise. If niches manage to skirt parti-
smaller units. Furthermore, given the institution- sanship and be valued by (but not beholden
building benefits of recruiting people who to) successive political appointees, political
already have proto-bureaucratic orientations, elites atop the organization may facilitate the
niches may be more likely to exist in sectors strategic diffusion of personnel, as in the
corresponding to particularly sought-after or example of PARD employees moving over to
effective university programs, or sectors the Budget Division. Yet the very conditions
where people have significant foreign educa- that enable interstitial bureaucracy may also
tion or work experience. Country-specific limit the ability to scale up. Interstitial bureau-
patterns in university or international educa- cracies feed off the mediocrity of others to
tion may therefore be correlated with the fuel their own hyper-performance and
location of bureaucratic niches. enhance in-group solidarity through a sense
This analysis also raises hypotheses about of relative superiority and pride. This ques-
conditions under which interstitial bureau- tions the idea that bureaucratic interstices will
cracy may spread, be overwhelmed, or—often be the spark that spreads bureaucracy like
overlooked—reach a durable equilibrium as wildfire throughout the state.
an interstice. We might hypothesize that the For organizational sociologists, this article
greater the dependency on inputs from exter- updates canonical bureaucratic theory, high-
nal sources (especially vast or varied sources), lighting structural vulnerabilities and organi-
the more likely fledgling niches are to be zational responses of interstitial bureaucracy
overwhelmed, resulting in demise. Con- not present in bureaucracy’s hegemonic form.
versely, self-contained functions—like the It advances understanding of large formal
FSM, which collect all their own data—may organizations as simultaneously unitary and
be less likely to be overwhelmed. Self-con- composed of many more-or-less autonomous
tainment may also constrain diffusion, result- parts, which are still part of the whole. This
ing in durable interstitiality. Beyond upstream inherent tension in many large formal organi-
dependencies, niches that directly serve the zations is important and under-theorized in
public may leverage public reputations for organizational research, where dominant the-
excellence into a support coalition, which can ories treat organizations as either coherent
506 American Sociological Review 82(3)

actors (e.g., principal-agent theory) or loose might have a disproportionate amount of aid
networks (e.g., Davis 2009). This work sug- directed to health, increasing the effectiveness
gests that most organizations, including of each dollar spent thanks to the enhanced
nation-states, have properties of both, a dual- competence of the local development partner.
ity that affects organizational practice. U.S. Bureaucracy is taken-for-granted—even
corporations likewise contain subunits that scorned as red tape—in the West. Advanced
distinctively outperform the rest of the corpo- industrialized nations are so accustomed to
ration (Sutton and Rao 2014). Ethnography Weberian bureaucracy that it appears to func-
shows high-performing corporate subunits tion like clockwork without a keeper. Intersti-
are similarly semi-autonomous and character- tial bureaucracy highlights that in some
ized by a distinctive ethos and strong insider contexts, the bureaucratic ethos is a fragile and
pride (Kunda 2009). This suggests dynamics deviant active project entangled with dominant
of interstitiality may generalize to high- local conditions. It thereby counters the logic
performing niches beyond interstitial bureau- of inevitability that sometimes pervades
cracy in neopatrimonial settings. accounts of institutional change. Too often, we
Interstitial bureaucracy raises policy impli- assume rationality and modernity are natural,
cations for the billion-dollar development overlooking the effort and commitment it took
industry. By highlighting how much the par- to foster them. Unlike a sweeping disciplinary
ticular form an adaptation takes depends on revolution, interstitial bureaucracy is a thou-
local context, it questions the efficacy of one- sand small revolutions quietly blooming in
size-fits-all global best practices. Capacity- rugged and unruly meadows.
building reforms should instead demonstrate
greater humility about the limitations of West- Acknowledgments
ern organizational wisdom. Reformers should
eagerly support and learn from local actors in Special thanks to Art Stinchcombe, who taught me to love
sociological approaches to organizations and thoughtfully
identified bureaucratic interstices, whose dual
commented on the earliest versions of this project. Thanks
cultural competency in the local environment to Jim Mahoney, whose insight shaped the project from the
and functional bureaucracy has enabled the beginning. I am deeply grateful to the ASR Deputy Editor
cultivation of practices with already-proven who handled my submission, my anonymous reviewers,
efficacy in that context. Donors might also and the Editors. They engaged the paper deeply, showcas-
ing the best of what a developmental review can be; this
rethink having foreign consultants fly on-site
paper has improved considerably due to the exchange with
for short training workshops intended to these anonymous but intellectually generous scholars.
enhance organizational capacity, which con- Thanks to other scholars who thoughtfully gave feedback
serves costs but yields fleeting or no improve- on this project and article, including Charles Camic, Bruce
ments. Given the importance of lived experience Carruthers, Nitsan Chorev, Jessica Collett, Michaela
DeSoucey, Barry Eidlin, Wendy Griswold, Geoff Hark-
for fostering the cognitive foundations of doing
ness, Carol Heimer, Jennifer Jones, Abigail Jorgensen,
bureaucracy, donors might instead fund Amy Langenkamp, John Levi Martin, Damon Mayrl,
medium-term international training experi- Elizabeth McClintock, Terence McDonnell, Matthew
ences for a critical mass of public servants from McEwen, Daniel Morrison, Abi Ocobock, Paul Ocobock,
the same department. This deep experience for Erin Rehel, Andrew Schrank, Jesper Sorensen, and Ann
Swidler. At early stages the project benefitted from feed-
a clustered critical mass is more likely to yield
back from the Northwestern University Comparative His-
durable change. Apart from capacity-building torical Social Science workshop and the UW Madison
aid, interstitial bureaucracy might occasion a Sociology of Economic Change and Development
rethinking of how public-facing development Research Seminar; the paper also benefitted from presen-
projects are distributed. Donors might system- tations at the Watson Institute at Brown University, the
Sociology Colloquium at Duke University, the University
atically target programmatic development aid
of Chicago Social Theory and Evidence Workshop, the
to sectors related to existing niches of excel- Stanford Graduate School of Business Organizational
lence within recipient states. For example, a Behavior Seminar, and the University of Notre Dame
state with a highly effective health ministry Culture Workshop and Africa Working Group.
McDonnell 507

Funding   9. Grindle (1997) likewise identifies personnel auton-


omy as a key characteristic distinguishing relatively
The fieldwork for this project would not have been pos-
high-performing agencies from lackluster ones.
sible without generous funding from the National Sci-
10. This should not be read as nothing-but-profession-
ence Foundation (Doctoral Dissertation Improvement
alization. Economists, lawyers, and engineers—
Grant #0728059), as well as research fellowships from
professional identities affiliated with the four
Northwestern University’s Department of Sociology and
niches—exist throughout public service and are not
Program of African Studies. This project has also benefit-
everywhere as successful where the rest of the inter-
ted from a Weinberg Dissertation Writing Fellowship
stitial mechanics are not present.
from Northwestern University.
11. See the social psychology on optimal distinctive-
ness (Brewer 1991).
Editors’ Note 12. Methodological limitations include, for example,
that contemporaneously observing interstitial
To avoid any possible conflict of interest, the entire
bureaucratic “birth” would require observing myr-
review process for this article was handled by an ASR
iad government agencies hoping one becomes a rare
Deputy Editor with no Notre Dame affiliation.
bureaucratic interstice during fieldwork—a costly
and potentially fruitless endeavor.
13. Interest in where within the state niches are located
Notes
may be grounded in an intuition that government
  1. Some scholars use “bureaucracy” to mean “govern- function must matter (see the earlier discussion on
ment administrative structure” (e.g., “representative function-based explanations). Systematically iden-
bureaucracy”). Because I argue that not all states tifying niche-location patterns is hampered by the
are bureaucratic in the Weberian sense, hereafter lack of cross-national comparative data on within-
“bureaucracy” refers to Weberian-style bureau- state measures of organizational capacity.
cratic ethos, and “administration” denotes a formal
organizational system in general.
  2. Neopatrimonialism is a system in which elites use References
state resources to enrich themselves and cultivate Abbott, Andrew. 1988. The System of Professions: An
loyal supporters through money, in-kind benefits, Essay on the Division of Labor. Chicago: University
or patronage employment. It is a particularized of Chicago Press.
network-based site and logic governing important Acemoglu, Daron, Simon Johnson, and James A. Robinson.
decisions decoupled from office or generalized col- 2005. “Institutions as a Fundamental Cause of Long-Run
lective good (Cammack 2007). Growth.” Handbook of Economic Growth 1:385–472.
  3. By effective, I refer to organizations, relative to oth- Adams, Julia. 2005. The Familial State. Ithaca, NY: Cor-
ers in their environment, with a widely shared repu- nell University Press.
tation for accomplishing organizational goals and Adams, Julia, Elisabeth S. Clemens, and Ann Shola Orl-
doing so in a largely impartial manner serving the off, eds. 2005. Remaking Modernity. Durham, NC:
public interest. This article does not engage debates Duke University Press.
about effectiveness versus efficiency. Similarly, Anderson, Elijah. 2000. Code of the Street: Decency, Vio-
predatory, clientelist, or corrupt states may also lence, and the Moral Life of the Inner City. New York:
operate well-managed projects that are not public- W. W. Norton & Company.
oriented, but that is not the focus of this article. Andrews, Matt, Lant Pritchett, and Michael Woolcock.
  4. Venue shopping is the process of policy supporters 2013. “Escaping Capability Traps through Problem-
seeking a more receptive location or legislative com- Driven Iterative Adaptation (PDIA).” World Develop-
mittee in which to push proposed policy changes. ment 51:234–44.
 5. Generally, high bureaucratic effectiveness or effi- Aylmer, G. E. 1961. The King’s Servants: The Civil Ser-
ciency serves as a foundation to enable high state vice of Charles I, 1625–1642. London, UK: Rout-
capacity. However, some scholars argue that the Chi- ledge & Kegan Paul.
nese state demonstrates high state capacity absent Barma, Naazneen H., Elisabeth Huybens, and Lorena
high bureaucratic efficiency (Zhou et al. 2012). Viñuela. 2014. Institutions Taking Root: Building
  6. This might also be considered a specialized division State Capacity in Challenging Contexts. Washington,
of labor. DC: World Bank Publications.
  7. This relates to managing boundary-spanning inter- Bates, Robert H. 1974. “Ethnic Competition and Mod-
dependencies and environmental uncertainty from ernization in Contemporary Africa.” Comparative
classic organizational theory (see Thompson 1967). Political Studies 6(4):457–84.
  8. Ghanaian Pidgin is simplified English coupled with Bates, Robert H. 1981. Markets and States in Tropical
local expressive conventions. For example, Ghanaian Africa: The Political Basis of Agricultural Policies.
pidgin when departing is “Eee Chale, I da go come.” Berkeley: University of California Press.
Chale is an expressive convention similar to “chap.” Baumgartner, Frank R., and Bryan D. Jones. 1993. Agen-
“I da go come” is the literal translation of the Twi “Me das and Instability in American Politics. Chicago:
kō na ba” (I go and come), used when departing. University of Chicago Press.
508 American Sociological Review 82(3)

Binder, Amy. 2007. “For Love and Money: Organiza- Ertman, Thomas. 1997. Birth of the Leviathan. Cam-
tions’ Creative Responses to Multiple Environmental bridge, UK: Cambridge University Press.
Logics.” Theory and Society 36(6):547–71. Evans, Peter. 1989. “Predatory, Developmental, and
Blattman, Christopher, and Laura Ralston. 2015. “Gen- Other Apparatuses: A Comparative Political Econ-
erating Employment in Poor and Fragile States: omy Perspective on the Third World State.” Socio-
Evidence from Labor Market and Entrepreneurship logical Forum 4(4):561–87.
Programs.” Available at SSRN. Evans, Peter. 1995. Embedded Autonomy: States and
Bratton, Michael, and Mwangi S. Kimenyi. 2008. “Vot- Industrial Transformation. Princeton, NJ: Princeton
ing in Kenya: Putting Ethnicity in Perspective.” Jour- University Press.
nal of Eastern African Studies 2(2):272–89. Evans, Peter, and James E. Rauch. 1999. “Bureaucracy
Brewer, Marilynn B. 1991. “The Social Self: On Being and Growth: A Cross-National Analysis of the Effects
the Same and Different at the Same Time.” Personal- of ‘Weberian’ State Structures on Economic Growth.”
ity and Social Psychology Bulletin 17(5):475–82. American Sociological Review 64(5):748–65.
Bromley, Patricia, and Walter W. Powell. 2012. “From Evans, Peter B., Dietrich Rueschemeyer, and Theda
Smoke and Mirrors to Walking the Talk: Decoupling Skocpol, eds. 1985. Bringing the State Back In. New
in the Contemporary World.” Academy of Manage- York: Cambridge University Press.
ment Annals 6(1):483–530. Evrensel, Ayşe Y. 2010. “Corruption, Growth, and
Brooks, Wyatt, Kevin Donovan, and Terence R. Johnson. Growth Volatility.” International Review of Econom-
2015. “Local Knowledge and Managerial Capital: ics & Finance 19(3):501–514.
Evidence from Kenyan Microenterprises.” Working Ferreira, Carlos, Michael Engelschalk, and William May-
Paper, University of Notre Dame, Notre Dame, IN. ville. 2007. “The Challenge of Combating Corrup-
Bruhn, Miriam, and Bilal Zia. 2013. “Stimulating Mana- tion in Customs Administrations.” Pp. 367–86 in The
gerial Capital in Emerging Markets: The Impact of Many Faces of Corruption: Tracking Vulnerabilities
Business Training for Young Entrepreneurs.” Journal at the Sector Level, edited by J. E. Campos and S.
of Development Effectiveness 5(2):232–66. Pradhan. Washington, DC: World Bank Publications.
Cammack, Diana. 2007. “The Logic of African Neopat- Fischer, Claude S. 1975. “Toward a Subcultural The-
rimonialism: What Role for Donors?” Development ory of Urbanism.” American Journal of Sociology
Policy Review 25(5):599–614. 80(6):1319–41.
Carpenter, Daniel P. 2001. The Forging of Bureaucratic Flatters, Frank, and W. Bentley Macleod. 1995. “Admin-
Autonomy: Reputations, Networks, and Policy Inno- istrative Corruption and Taxation.” International Tax
vation in Executive Agencies, 1862–1928. Princeton, and Public Finance 2(3):397–417.
NJ: Princeton University Press. Fligstein, Neil. 1987. “The Intraorganizational Power
Carruthers, Bruce G. 1994. “When Is the State Autono- Struggle: Rise of Finance Personnel to Top Leader-
mous? Culture, Organization Theory, and the Politi- ship in Large Corporations, 1919–1979.” American
cal Sociology of the State.” Sociological Theory Sociological Review 52(1):44–58.
12(1):19–44. Fourcade-Gourinchas, Marion, and Sarah L. Babb. 2002.
Centeno, Miguel A. 1997. “Blood and Debt: War and “The Rebirth of the Liberal Creed: Paths to Neoliber-
Taxation in Nineteenth-Century Latin America.” alism in Four Countries.” American Journal of Soci-
American Journal of Sociology 102(6):1565–1605. ology 108(3):533–79.
Chibber, Vivek. 2002. “Bureaucratic Rationality and the Gilbert, Jess, and Carolyn Howe. 1991. “Beyond ‘State
Developmental State.” American Journal of Sociol- vs. Society’: Theories of the State and New Deal
ogy 107(4):951–89. Agricultural Policies.” American Sociological Review
Clemens, Elisabeth S., and James M. Cook. 1999. “Poli- 56(2):204–220.
tics and Institutionalism: Explaining Durability and Giné, Xavier, and Ghazala Mansuri. 2014. “Money or
Change.” Annual Review of Sociology 25:441–66. Ideas? A Field Experiment on Constraints to Entre-
Davis, Gerald F. 2009. Managed by the Markets: How preneurship in Rural Pakistan.” June 1, World Bank
Finance Re-Shaped America. Oxford, UK: Oxford Policy Research Working Paper (6959).
University Press. Gorski, Philip. 2003. The Disciplinary Revolution: Cal-
DiMaggio, Paul J., and Walter W. Powell. 1983. “The vinism and the Rise of the State in Early Modern
Iron Cage Revisited: Institutional Isomorphism and Europe. Chicago: University of Chicago Press.
Collective Rationality in Organizational Fields.” Grindle, Merilee S. 1997. “Divergent Cultures? When
American Sociological Review 48(2):147–60. Public Organizations Perform Well in Developing
Dobbin, Frank. 1994. Forging Industrial Policy: The Countries.” World Development 25(4):481–95.
United States, Britain, and France in the Railway Grosh, Barbara. 1991. Public Enterprise in Kenya: What
Age. New York: Cambridge University Press. Works, What Doesn’t, and Why. Boulder, CO: Lynne
du Gay, Paul. 2005. The Values of Bureaucracy. London, Rienner Publishers.
UK: Oxford University Press. Guthrie, Doug. 2001. Dragon in a Three-Piece Suit: The
Eliasoph, Nina, and Paul Lichterman. 2003. “Culture in Inter- Emergence of Capitalism in China. Princeton, NJ:
action.” American Journal of Sociology 108(4):735–94. Princeton University Press.
McDonnell 509

Hallett, Tim, and Marc J. Ventresca. 2006. “Inhabited Levi, Margaret. 1988. Of Rule and Revenue. Berkeley:
Institutions: Social Interactions and Organizational University of California Press.
Forms in Gouldner’s Patterns of Industrial Bureau- Lizardo, Omar. 2017. “Improving Cultural Analysis:
cracy.” Theory and Society 35(2):213–36. Considering Personal Culture in its Declarative and
Heimer, Carol A. 1999. “Competing Institutions: Law, Nondeclarative Modes.” American Sociological
Medicine, and Family in Neonatal Intensive Care.” Review 82(1):88–115.
Law and Society Review 33(1):17–66. Lizardo, Omar, and Michael Strand. 2010. “Skills, Tool-
Helman, Gerald B., and Steven R. Ratner. 1992. “Saving kits, Contexts and Institutions: Clarifying the Rela-
Failed States.” Foreign Policy Winter:3–20. tionship between Different Approaches to Cognition
Henisz, Witold J., Bennet A. Zelner, and Mauro F. Guil- in Cultural Sociology.” Poetics 38:204–27.
lén. 2005. “The Worldwide Diffusion of Market- Lofland, John, and Lynn Lofland. 1995. Analyzing Social
Oriented Infrastructure Reform, 1977–1999.” Ameri- Settings: A Guide to Qualitative Observation and
can Sociological Review 70(6):871–97. Analysis. New York: Wadsworth.
Hooks, Gregory. 1990. “The Rise of the Pentagon Macaulay, Stewart. 1963. “Non-Contractual Relations in
and U.S. State Building: The Defense Program as Business: A Preliminary Study.” American Sociologi-
Industrial Policy.” American Journal of Sociology cal Review 28(1):55–67.
96(2):358–404. Mann, Michael. 1986. The Sources of Social Power. Vol.
Huber, John D., and Nolan McCarty. 2004. “Bureaucratic 1, A History of Power from the Beginning to Ad 1760.
Capacity, Delegation, and Political Reform.” Ameri- New York: Cambridge University Press.
can Political Science Review 98(3):481–94. Mauro, Paolo. 1995. “Corruption and Growth.” Quar-
Jepperson, Ronald L., and John W. Meyer. 1991. “The terly Journal of Economics 110(3):681–712.
Public Order and the Construction of Formal Organi- McKenzie, David, and Christopher Woodruff. 2013.
zations.” Pp. 204–231 in The New Institutionalism in “What Are We Learning from Business Training and
Organizational Analysis, edited by W. W. Powell and P. J. Entrepreneurship Evaluations around the Developing
DiMaggio. Chicago: The University of Chicago Press. World?” World Bank Research Observer 29(1):48–82.
Johnson, Martha C. 2015. “Donor Requirements and Meyer, John W., John Boli, George M. Thomas, and
Pockets of Effectiveness in Senegal’s Bureaucracy.” Francisco O. Ramirez. 1997. “World Society and
Development Policy Review 33(6):783–804. the Nation-State.” American Journal of Sociology
Karlan, Dean, and Martin Valdivia. 2011. “Teaching 103(1):144–81.
Entrepreneurship: Impact of Business Training on Meyer, John W., and Brian Rowan. 1977. “Institution-
Microfinance Clients and Institutions.” Review of alized Organizations: Formal Structure as Myth
Economics and Statistics 93(2):510–27. and Ceremony.” American Journal of Sociology
Kaufmann, Daniel, Aart Kraay, and Massimo Mastru- 83(2):340–63.
zzi. 2010. “Worldwide Governance Indicators.” The Migdal, Joel S. 2001. State in Society: Studying How
World Bank (http://data.worldbank.org/). States and Societies Transform and Constitute One
Kenya, Colony and Protectorate of. 1960. Statistical Another. New York: Cambridge University Press.
Abstract. East African Statistical Department-Kenya Morgan, Kimberly J., and Ann Shola Orloff, eds. 2016.
Unit. Nairobi: The Government Printer. The Many Hands of the State: Theorizing Political
Knack, Stephen, and Philip Keefer. 1995. “Institutions Authority and Social Control. New York: Cambridge
and Economic Performance: Cross-Country Tests University Press.
Using Alternative Institutional Measures.” Econom- O’Donnell, Guillermo A. 2004. “Why the Rule of Law
ics & Politics 7(3):207–227. Matters.” Journal of Democracy 15(4):32–46.
Kohli, Atul. 2004. State-Directed Development: Political Parkinson, C. Northcote. 1955. “Parkinson’s Law.” The
Power and Industrialization in the Global Periphery Economist, November 19.
New York: Cambridge University Press Pogoson, Aituaje Irene, and Michael Roll. 2014. “Turning
Kunda, Gideon. 2009. Engineering Culture: Control and Nigeria’s Drug Sector Around: The National Agency
Commitment in a High-Tech Corporation. Philadel- for Food and Drug Administration and Control (NAF-
phia: Temple University Press. DAC).” Pp. 97–127 in The Politics of Public Sector
La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Performance, edited by M. Roll. New York: Routledge.
Shleifer, and Robert Vishny. 1999. “The Quality of Quadagno, Jill. 1987. “Theories of the Welfare State.”
Government.” Journal of Law, Economics, and Orga- Annual Review of Sociology 13:109–128.
nization 15(1):222–79. Rauch, James E., and Peter B. Evans. 1997. “Bureau-
Leonard, David K. 1991. African Successes: Four Public cratic Structure and Economic Performance” (http://
Managers of Kenyan Rural Development. Berkeley: weber.ucsd.edu/~jrauch/research_bureaucracy.html).
University of California Press. Roll, Michael, ed. 2014. The Politics of Public Sector
Leonard, David K. 2010. “‘Pockets’ of Effective Agen- Performance: Pockets of Effectiveness in Developing
cies in Weak Governance States: Where Are They Countries. New York: Routledge.
Likely and Why Does It Matter?” Public Administra- Rueschemeyer, Dietrich. 1986. Power and the Division
tion and Development 30(2):91–101. of Labour. Stanford, CA: Stanford University Press.
510 American Sociological Review 82(3)

Rueschemeyer, Dietrich. 2005. “Building States—Inher- Weber, Klaus, Gerald F. Davis, and Michael Lounsbury.
ently a Long-Term Process? An Argument from The- 2009. “Policy as Myth and Ceremony? The Global
ory.” Pp. 143–64 in States and Development, edited Spread of Stock Exchanges, 1980–2005.” Academy
by M. Lange and D. Rueschemeyer. London, UK: of Management Journal 52(6):1319–47.
Palgrave Macmillan Weber, Max. [1909] 1956. “Speech to Verein Für Sozi-
Schrank, Andrew. 2016. “Imported Institutions: Boon or alpolitik.” Pp. 125–31 in Max Weber and German
Bane in the Developing World?” Presented at “Weak Politics: A Study in Political Sociology, edited by J. P.
Institutions in Latin America: New Theoretical and Mayer. London, UK: Faber and Faber Ltd.
Empirical Approaches,” May 9–10, Harvard Univer- Weber, Max. 1978. Economy and Society: An Outline
sity, Cambridge, MA. of Interpretive Sociology, edited by G. Roth and C.
Scott, James C. 1998. Seeing Like a State. New Haven, Wittich. Translated by G. Roth. Berkeley: University
CT: Yale University Press. of California Press.
Selznick, Philip. 1949. TVA and the Grass Roots: A Study White, Leonard. 1954. The Jacksonians: A Study of Admin-
of Politics and Organization. London, UK: Cam- istrative History: 1829–1861. New York: Macmillan.
bridge University Press. Wohl, Hannah. 2015. “Community Sense: The Cohesive
Singh, Prerna. 2015. “Subnationalism and Social Devel- Power of Aesthetic Judgment.” Sociological Theory
opment: A Comparative Analysis of Indian States.” 33(4):299–326.
World Politics 67(3):506–562. Woo-Cumings, Meredith, ed. 1999. The Developmental
Sorensen, Jesper. 2015. “Personal Communication to State. Ithaca, NY: Cornell University Press.
Author.” Palo Alto, CA. World Bank. 1997. “World Development Report 1997:
Steinmetz, George, ed. 1999. State/Culture: State-For- The State in a Changing World.” Washington, DC:
mation after the Cultural Turn. Ithaca, NY: Cornell World Bank.
University Press. World Bank. 1999–2007. “World Bank Governance and
Strauss, Julia C. 1998. Strong Institutions in Weak Poli- Anti-Corruption (GAC) Diagnostic Surveys” (http://
ties: State Building in Republican China, 1927–1940. www.gaportal.org/resources/detail/world-bank-gov
Oxford, UK: Oxford University Press. ernance-and-anti-corruption-gac-diagnostic-surveys).
Sutton, Robert I., and Hayagreeva Rao. 2014. Scaling Up World Bank. 2002. “World Development Report 2002:
Excellence: Getting to More without Settling for Less. Building Institutions for Markets.” Washington, DC:
New York: Crown Business. World Bank.
Swidler, Ann. 1986. “Culture in Action: Symbols World Bank. 2003. “World Development Report 2003:
and Strategies.” American Sociological Review Sustainable Development in a Dynamic World.”
51(2):273–86. Washington, DC: World Bank.
Swidler, Ann. 2016. “Personal Communication.” Seattle, World Bank. 2015. “World Development Report 2017:
WA. Governance and the Law.” Washington, DC: World
Tendler, Judith. 1997. Good Government in the Tropics. Bank (http://go.worldbank.org/LBM5VYIFI0).
Baltimore, MD: Johns Hopkins University Press. World Bank. 2016. “Projects & Operations by Fiscal
Thompson, James D. 1967. Organizations in Action: Year.” Washington, DC. Retrieved January 4, 2016
Social Science Bases of Administrative Theory. New (http://www.worldbank.org/projects/search?lang=en
York: McGraw-Hill. &searchTerm=&sectorcode_exact=BC).
Tilcsik, András. 2010. “From Ritual to Reality: Demogra- Zhou, Xueguang, Yun Ai, and Hong Lian. 2012. “The
phy, Ideology, and Decoupling in a Post-Communist Limit of Bureaucratic Power in Organizations: The
Government Agency.” Academy of Management Case of the Chinese Bureaucracy.” Pp. 81–111 in
Journal 53(6):1474–98. Rethinking Power in Organizations, Institutions, and
Tilly, Charles. 1990. Coercion, Capital, and European Markets, edited by D. Courpasson, D. Golsorkhi, and
States, AD 990–1990. Cambridge, MA: Basil Blackwell. J. Sallaz. Bingley, UK: Emerald Group Publishing
Transparency International. 2010. “Global Corruption Limited.
Barometer: Tabulations by Country.” Berlin, Ger-
many: Transparency International. Erin Metz McDonnell is a Kellogg Assistant Professor
Uzzi, Brian, and Jarrett Spiro. 2005. “Collaboration and of Sociology at the University of Notre Dame. Her
Creativity: The Small World Problem.” American research cuts across organizational, political, cultural,
Journal of Sociology 111(2):447–504. and economic sociology, focusing on how social organi-
Vaisey, Stephen. 2009. “Motivation and Justification: A zation affects economic outcomes. Her work on Webe-
Dual-Process Model of Culture in Action.” American rian budgetary units, published in the American Journal
Journal of Sociology 114(6):1675–1715. of Sociology, won the 2015 Theory Award. Her current
Vedres, Balazs, and David Stark. 2010. “Structural Folds: work intersects Weberian bureaucracy and interstitial
Generative Disruption in Overlapping Groups.” organizational theory to understand sub-state variation in
American Journal of Sociology 115(4):1150–90. administrative capacity in the global South.

You might also like