I think you [Peter G. Stillman] are thinking in terms of the common
misintepretation of Coase. Coase's main point was not
that all is well when transaction costs are low.
Coase's point was that the notion of zero or neglible
transaction costs is utterly unrealistic. The best
institutions, to COase, are the ones that result in
the lowest possible costs to effecting welfare
enhancing exchanges. Transaction costs are always
significant, so externalities always exist and social
costs diverge from private costs. But this alone
proves nothing. THe costs of effecting worthwhile
exchanges (or transfers) through government mean that
government policies will also deviate from first best
optimality. So the relevant choice is not between
nearly perfect markets (with negligible transaction
costs, among other things) and government as we
know it. Nor is it between imperfect markets and a
nirvana of perfect government, run by benevolent
omniscient dictators. The real choice is between
private and public sector institutions, none of which
can ever come close to delivering first best pareto
optimality. So one lesson from Coase is not to be
naive about any institution delivering perfection, and
not to hold any institution to the absurd standard of
first best competitive equilibrium. 

Another lesson from Coase (and Hayek) is that real
errors exist, and we never find ourselves in the best
of all possible worlds. I have heard some economists
talk about the stuff I wrote about in the first
paragraph, and then conclude that we always end up
with the second best institutions. Coase denied this
explicitly. I would not characterize COase as an
equilibrium theorist. Coase saw institutions as
changing through time, for better or worse. Coase also
disparaged mainstream 'blackboard economics'. 

I have a few problems with Coase, but the points you
raise derive from widely held misconceptions of
Coase's work, in my opinion.

Doug Mackenzie