> There is no obvious inherent reason why a government would even be able to estimate externalities better than a market
Well, assuming effective equal democracy [1] there is, given also the assumptions on which market efficiency in transactions without externalities is based [2], since presumably interest in promoting or preventing a transaction one is not engaged in directly engaged maps to the expected utility or disutility experienced as a result of the transaction (i.e., its positive or negative externality.)
[1] which is a big assumption, to be sure.
[2] which are also big assumptions, but they are also the entire underpinning for the idea that libertarian, market-based economies are efficient for any kind of transaction, so in considering regulated vs. libertarian markets, its pretty fair to assume them in consideration on the regulated side.
Indeed, all the economic models I'm aware of are explicitly theoretical, including ones that cover how governments and markets work. They often contain assumptions that are obviously not present in the real world: like perfect information or zero transaction costs. Of course, that doesn't mean they're not occasionally testable to some extent, but I'm strongly skeptical of anyone claiming to have a "proof" of any nontrivial claim regarding these ideas. That said, I do personally find the arguments that markets will outperform governments more convincing than the inverse.
On the topic of market transactions with externalities, you might find the Coase theorem interesting, and the related folk theorems in game theory. I see it as a compelling rebuttal to the Pigovian model of using a centralized organization (generally government) to internalize social costs.
> On the topic of market transactions with externalities, you might find the Coase theorem interesting, and the related folk theorems in game theory. I see it as a compelling rebuttal to the Pigovian model of using a centralized organization (generally government) to internalize social costs.
One can envision idealized democratic government as a mechanism for realizing the required precondition for the Coase theorem, to wit, the democratic process of policy making in such a system is the venue in which externalities are traded. (That such exchanges occur in the policy making process in any government of human beings is widely observed, the "ideal democratic" part is that that all citizens have equal power in the policy making process, such that the trade is generalized.)
This, of course, unifies the Coase theorem with the Pigovian model (the separation between the two comes from treating "government" as something distinct from the participants in the economy, rather than a venue in which those participants interact.)
Well, assuming effective equal democracy [1] there is, given also the assumptions on which market efficiency in transactions without externalities is based [2], since presumably interest in promoting or preventing a transaction one is not engaged in directly engaged maps to the expected utility or disutility experienced as a result of the transaction (i.e., its positive or negative externality.)
[1] which is a big assumption, to be sure.
[2] which are also big assumptions, but they are also the entire underpinning for the idea that libertarian, market-based economies are efficient for any kind of transaction, so in considering regulated vs. libertarian markets, its pretty fair to assume them in consideration on the regulated side.