A great post by Felix Salmon discusses externalities, congestion pricing, and a spreadsheet by Charles Komanoff.  The comparison of the way the resulting costs from congestion pricing fall, depending on which scheme is used, is particularly important, as the original NYC plan would have disproportionately hit middle-class commuters from Brooklyn and Queens.  Congestion pricing is an excellent example of tactically hacking a vast and complex existing system; while it requires a lot of hard and rigorous work to do correctly, it is exactly the sort of strategy that can be used to transform conditions in a megacity with a great deal of infrastructural inertia.