state agencies, adapt new purposes and continue to exist. Even oddities like the
electoral college in the USA or an ‘‘executive second chamber’’ in Germany from
Bismarck to Adenauer have not been changed in spite of numerous reform
initiatives. Even the occupation forces in Germany after 1945 failed in trying to
impose on West Germany diVerent systems of a federal chamber, diVerent forms
of industrial relations, or a uniWed social security system. The cold war soon
promoted other priorities than the overhauling of traditional institutions.
Organizational theory has developed many strategies for the reform of political
institutions. They were more successful in the revived ‘‘new institutional econom-
ics’’ in the context of enterprises and industrial relations (Richter 1994 , 3 ). The ‘‘new
institutionalism’’ in political science, however, has to live with the fact of the
persistence of many forms of organizational routines and structures. Most
institutional reform proved to be ad hoc activity (March and Olsen 1989 ,69V).
There is a permanent division in political science between the ‘‘hard’’ type of
analysis aiming at universal laws—as in behavioralism and rational choice—and
the ‘‘soft’’ historically oriented analysis of political events and lines of cultural
development. The hope remains that both camps engage in a fruitful exchange
(Rothstein 1996 , 156 ). Thenew institutionalismwas a major step in the direction of
this synthesis. March and Olsen ( 1984 , 747 ) hoped for a ‘‘gentle confrontation
between the wise and the smart’’ which characterizes innovations in intellectual
history. Many movements and theories have called themselves ‘‘new.’’ As in other
Welds—such as art—they quickly ended in ‘‘post-’’movements. In the best case
this lead to a development ‘‘from post- to neo.’’ Is neoinstitutionalism really new?
(a) It diVers from the older institutionalism in the attempt to work theory-
oriented. (b) It contains the achievements of former revolts—such as the behav-
ioral and rational choice revolts—to diVerentiate between dependent and
independent variables, though some authors blur this diVerence and treat their
institutions simultaneously as dependent and independent (Pedersen 1991 , 131 f).
Neoinstitutional approaches observe actual behavior instead of legal and formal
aspects of political behavior which prevailed in older theories. (c) The main virtue
is that concepts have been developed which make new institutionalism more
comparative than the older juxtapositions of regimes in early institutionalism
(Peters 1996 , 206 ).
Comparative studies on institutions in Europe developed between European
traditions and American innovations. TheWrst foreign inXuences on my own
thinking took place in France in the late 1950 s. As a student in France, Duverger
and Aron exercised considerable inXuence. My book onPolitical Parties in Western
Democracies( 1985 ) has sometimes been dubbed as an ‘‘updated version’’ of Duver-
ger’s study. This perception hardly did justice to my own intentions:
comparative studies of institutions according to my interests had to get rid of
three vices of the older institutionalism in France: (a) The preoccupation with a
unilinear causality between electoral laws and parties in the school of Andre ́
Siegfried and Duverger; (b) The benign neglect for foreign languages besides
political institutions—old and new 753