Much of the modern complexity movement in economics is stimulated by
the use of more current models out of physics, especially those drawing on
statistical mechanics.  These are especially useful for studying systems 
with
heterogeneous interacting agents.  The approach is generally not to prove
theorems in a Bourbakist manner, but to engage in simulations or to try to
fit various data from financial markets (notorious for their non-normal fat 
tails
in returns) or income or wealth or firm size distributions to various models
that come out of physics and generate power law distributions.
     Not sure if Deirdre approves of this sort of thing or not, but it is 
not
Bourbakist, and the econometrics is usually not so oriented to stat 
significance.
Much of this work is actually being done by physicists, now part of the 
so-called
"econophysics" movement, although sometimes the ignorance of economics by
these people can be annoying, especially when they start intoning about how 
they
have solved all the problems of economics, when they have not read much 
literature
and are talking about non-problems, although there seems to be less of this 
sort of
thing going on these days.

Barkley Rosser