answer reveals some odd features in the comparative analysis of cabinets and prime
ministers.
First, much of the published work on Westminster systems is not strictly
speaking comparative, but compilations of country studies (see Weller, Bakviss,
and Rhodes 1997 , 7 – 11 for citations). Nothing wrong with that, but it is not
comparative analysis (cf. Weller 1985 ). Second, the modernist-empiricist project
can take such labels as ‘‘cabinet’’ at face value and compare apples and oranges.
Switching cabinets for (say) ministers will not solve the problem (see Blondel and
Thie ́bault 1991 ). Third, most comparative research is between regime types, not on
variations within one regime type (see Shugart, Chapter 18 ). Finally, and poten-
tially the most misleading of all, there are the comparisons of American presidents
with UK prime minsters. Rose ( 2001 , 237 – 44 ) identiWes one similarity—the impact
of the mass medias in personalizing chief executives and election campaigns—and
many diVerences, including diVerent recruitment and career paths, direct
popular election of the president,Wxed term of oYce, constitutional checks and
balances, and limited control of the legislature and, therefore, domestic policy. It
might seem an overly simple-minded conclusion but the comparative analysis of
prime ministers and cabinets needs to compare like with like. It is simply not
revealing to be told there are big diVerences in the powers of prime ministers, there
are big diVerences in the powers of presidents, and there are big diVerences between
prime ministers and presidents.
There are two fruitful lines of analysis: rational choice institutionalism (see
Strøm, Mu ̈ller, and Bergman 2003 ) and core executive models (see Elgie 1997 ,
1998 ). I have provided already a brief summary of Strøm’s principal–agent theory
(above, p. 326 ). Alternatively, Elgie suggests we use the six models of core executive
operations to analyse prime ministerial and semi-presidential systems:
- Monocratic government—personal leadership by prime minister or president.
- Collective government—small, face-to-face groups decide with no single mem-
ber controlling.- Ministerial government—the political heads of major departments decide
policy.- Bureaucratic government—non-elected oYcials in government departments
and agencies decide policy. - Shared government—two or three individuals have joint and equal responsi-
bility for policy-making. - Segmented government—a sectoral division of labour among executive actors
with little or no cross-sectoral coordination.
The advantage of this formulation is that it gets away from bald assertions about
theWxed nature of executive politics. While only one model may operate at any one
time, there can still be aXuid pattern as one model succeeds another. It also
concentrates the mind on the questions of which model of executive politics
executives in parliamentary government 337