Davis, Lucas and Matthew Kahn. 2010. International trade in used vehicles: The environmental consquences of NAFTA. Economic Policy. 58-82.
Davis and Kahn set up a nice little set of models to help us understand what has likely happened in the durable goods market for vehicles. In comparatively higher income countries, used durables like cars are likely to get traded out to lower income countries–here, the US and Mexico. And since older durables emit more than new cars, they find that this robust trade in used vehicles increases lifetime emissions as Mexico consumers substitute away from transit use to used car consumption and those cars stay in use longer. An excellent paper: I highly encourage you to go read (and to spring for membership in the American Economic Association: you get lots of good journals and a calendar with economist centerfolds! One of my happiest investments this year.)
A couple of weak points: they say at the beginning that they establish that trade makes emissions go up in both countries. No, they actually show that emissions go down in the US but up in Mexico, and the increases in Mexico outstrip the reductions in the US. I don’t love the way they calculate emissions: they have to make some assumptions about the distribution of vehicle miles of travel, and I suspect that it is possible, given their analysis, that trade make makes VMT go up in both countries. Moreover, they note that costs of repairs are low in Mexico, yet they really don’t calculate how repairs can significantly improve engine performance. A car isn’t as good as new, but that doesn’t mean it stays a clunker after it’s traded. This may be particularly true depending on where the used car ends up in Mexico: Mexico City has different incentives and regulations for fixing up a car than other parts of the country.