Gravity modelling is Newtonian physics adapted for economic forecasting. Just as the attraction between two heavenly bodies is directly proportional to their masses and inversely proportional to the distance between them, so the volume of trade and the amount of foreign direct investment between two countries depends on how big and how geographically close they are.
Using this approach, the Treasury says there would be a 43% loss of trade with the EU were the UK to revert to WTO rules, and because almost half the UK’s trade is with the EU this would result in a 24% loss in total trade. The assumption is that there has been a 76% increase in UK trade as a result of membership of the EU and that all of these gains would be lost. There would be no gains in trade with non-EU countries to compensate for the loss.
The gravity model is the standard trade model. And I don’t think it’s right. As in an email I’ve just sent:
Small thought about the gravity model of trade.
Historically geographic closeness might have made sense as a measure. At some time that is. But if we cast further back not so much. Say, just to invent an example, trade over the Pennines was at one time much smaller than trade over the Irish Sea. Ship and river transport was much more important than road (there were no roads).
Or coal from Newcastle to London by ship but near nothing from Newcastle to Carlisle.
The gravity model should thus be tweaked to measure economic geography, which methods of transport are in use? Or, how close is somewhere in travel days, not miles? Travel cost perhaps not miles.
At which point much more makes sense. The container network will move 30 tonnes of anything anywhere for under $5,000 these days. Geographic proximity, what that gravity model assumes, is rather less important.
In fact, transport costs between Birmingham and Barcelona, Birmingham and Birmingham AL and Birmingham and Brisbane are not notably different these days.
Gravity as measured by trade costs rather than geography would make much, much, more sense.