Friday, May 7, 2010

Increasing trade by creating more borders

What is the difference between macroeconomics and international macroeconomics or trade? There is a border. Just splitting an economy in two seems trivial, yet it matters. A lot. There is plenty of empirical evidence that borders matter. They inhibit trade. They allow for purchasing-power-parity deviations to persist. The economic well-being can differ dramatically across a border despite geographic similarities.

Emmanuelle Lavallée and Vincent Vicard note that the number of borders has considerably increased since World War II, with the number of countries going from 72 to 192. Given the border effect, this should be bad. But this has also a important side effect: transactions that were internal become international, thus boosting international trade statistics. Lavallée and Vicard find that measured international trade has increased by 9% solely because of new borders, but actual trade would have been 4% higher without those borders. While this is not negligible, we need to keep in mind that world trade has increased by a factor of 30 during this period.

2 comments:

Unknown said...

I would be interested to see the evidence for borders actually mattering for trade and economic well-being.

I can certainly see why they might matter, but I'm inclined to believe that any simple analysis will dramatically overstate the impact of borders, because the erection of borders is endogenous to relevant economic factors.

That is to say, when we see South Korea and North Korea "create" a border, and observe the well-being of their peoples (or perhaps just one of their peoples) go down, I expect that it went down not because of the border, but because of the reason behind the use of the 38th parallel as a demarcating line.

Agent Continuum said...

The "border puzzle" may or may not be there: http://ideas.repec.org/p/cii/cepidt/2002-01.html

I guess that the literature on FTAs might have something to say about how to handle the endogeneity of barriers to trade, to some limited extent.