Hierarchy 77
operating core is especially infl uential, universities being the obvious example.
Th e divisionalized bureaucracy is commonly found in contexts in which outputs
need to be standardized but the need for services vary. State divisions of social
services are examples, and so is the Internal Revenue Service. Middle manage-
ment tends to be infl uential in such hierarchies. An adhocracy is the least formally
organized hierarchy; it tends to emphasize mutual adjustment and to engage in
team projects, to use matrix forms, and to mix centralization and decentraliza-
tion. Role clarity, sharp divisions of labor, chains of command, and standardiza-
tion are weak in adhocracies, but the search for innovation is strong. Computer
soft ware companies and the generalized R&D organizational format typify ad-
hocracies, which have become the ideal modern structure emphasizing limited
rules, fl exible time, entrepreneurial management, and customer service. In the
public sector, the NASA Manned Space Flight Center is oft en used as an illustra-
tion of adhocracy.
Any large-scale organization is likely to exhibit elements of each of these
forms, and eff ective managers tend to understand the linkages between alterna-
tive structural choices and likely results. Th ey know that structure matters, which
oft en explains the tendency of management to push reorganization. When that
happens, Mintzberg (1979, 1992), basing his study on private fi rms, suggests that
the several components of hierarchy will tend to pull in particular directions.
Th e strategic apex will pull in the direction of centralization and standardization.
Middle management will tend to balkanize and protect turf. Th e technostructure
will join the strategic apex in a pull to standardize. Th e support system will be in-
clined to collaborate and network. Finally, the operating core will see a powerful
pull to professionalize. Th e public-sector push to contract out and to privatize
does appear to run counter to the argument that the strategic apex will tend to
centralization and standardization. Although there is almost always resistance
at the operating core to contracting out, political pressure to downsize and save
money by doing so would appear to run counter to centralization. But contracts
are always replete with standards, and the contracting process may imply moving
elements of control away from middle management to the strategic apex, where
politically appointed persons are most infl uential. In contract regimes, the sup-
port staff ’s instinct to collaborate and network would appear to support contract-
ing on the assumption that contractors are new partners and that elements of
institutional structure and management can essentially be exported and hidden
(Light 1999).
Indeed, even though networked bureaucracy has garnered much attention,
and some think the trend is toward horizontal government (O’Toole 1997b;
Kettl 2002), others, such as Carolyn J. Hill and Laurence E. Lynn Jr. (2005),
argue conversely. Hill and Lynn contend that rather than supplanting hierar-
chical government, networked or horizontal governing is being added so as to
improve governance in an otherwise hierarchical system. Networks are impor-
tant and should be taken seriously, but they perhaps are not replacing traditional