this means increasing reliance on ‘‘soft law’’ (codes over commands); on modes
of self-regulation; and an emphasis precisely on ‘‘reXexivity’’—on malleability,
Xexibility, and a willingness to adapt and learn. 2 In short, lock-in and path
dependency arising from national historical experience are not determinate;
paths can change—and converge.
One important consequence of this account is to reinstateagencyas an inXuence
on institutional design. The possibilities are well illustrated in Ayres and
Braithwaite’s ( 1992 )inXuential model of enforced self-regulation. This attempts
to develop a theory of institutional choice, departing from a straightforward
universal emphasis on reXexivity: one where both choice in institutional design,
and choice of particular institutional instruments in particular regulatory circum-
stances, once again becomes a possibility. Part of the importance of agency in their
model rests on the notion that choices can be made between command and
reXexivity: in their world, regulatory authorities at the top of the regulatory
pyramid speak the soft language of reXexivity, but carry a big stick.
All these versions of reXexivity root institutional change in common structural
conditions across industrial societies, notably high social and technical complexity.
An alternative account is rooted in more contingent historical and institutional
circumstances. The best-known version is encapsulated in Majone’s theory of the
emergence of a new regulatory state ( 1991 , 1996 , 1999 ). This amounts to both an
empirical and a prescriptive theory of theconstitutionof economic life. Some of
Majone’s themes echo theorists of high complexity. This is particularly noticeable
in his argument that the regulation of economic life demands a Madisonian
constitution: a system that entrenches expert opinion and interested minorities
in the decision-making process, at the expense of modes of majoritarian constitu-
tions. Some of Majone’s arguments also respond to the political economy of
advanced capitalism after the end of the long boom, notably to the (alleged)
exhaustion of command modes in economic life associated with high Keynesian-
ism. Some respond more immediately to the problem of making sense of the
institutional forms being developed by, and appropriate for, the new system of
economic government developing in the European Union. All converge on
the claim that institutional structures have to display two features in the new
regulatory state: the state has to abstain from anything more ambitious than the
promulgation of broad rules governing the behavior of institutional actors in
economic life; and responsibility for the implementation of rules must be delegated
to the lowest possible institutional level. The latter, in practice, commonly means
institutional actors in markets—trade associations, standard setting institutes,
professional bodies, and individualWrms.
2 The convergence on reXexivity comes from very diVerent theoretical, and substantive, starting
points: I draw heavily on the theoretical work of Teubner 1987 , 1993 , and 1994 ; Collins 1999 on contract;
Gunningham, Grabosky, and Sinclair 1998 on environmental policy; and Gunningham and Johnstone
1999 on health and safety in the enterprise.
economic institutions 149