Appendix – Measuring Remoteness

If one looks at the globe, it is obvious that some countries are close to potential trading partners (eg, Germany) while other countries are in very remote locations (eg, New Zealand). Similarly, it is obvious that some country’s economic activity is concentrated in a relatively small area (Netherlands, Japan) while economic activity in other countries is quite dispersed (Canada, Australia). In this appendix, we propose a summary statistic to capture these elements of remoteness.

The remoteness measure adopted here comes from the literature on “Gravity Equations” in the international trade literature.[69] Consider the case of a town located on a road between two other towns, one on the left (town B) that is 100 kilometres away and one on the right (town C) that is 50 kilometres away. This literature asks – which town is closer, economically speaking, to town A? Or, to put this somewhat differently, if you opened a business in town A, how “important” would town B and town C be to your business plan?

The proposition in the literature on “Gravity Equations” suggests that the answer depends on the physical distance and the size of the town in economic terms. So if the amount of economic activity in town B (the one that is 100 kilometres away) is significantly larger (eg, ten times as much), an economic summary statistic tells us that the economic distance between towns A and B is less than the economic distance between towns B and C.

Conceptually, the basic scaling notion built into gravity models is that a town that is 100 miles away and has aggregate economic activity of $100 gets a 1.0. This is the same value that one gets for a town that is 50 miles away with aggregate economic activity of $50 as well as a town that is 200 miles away and has $200. This scaling is, of course, arbitrary, but it stands up as a reasonable approximation in the empirical literature.

For each country “i”, the remoteness index Ri is given by:

Ri = +

Where is the internal distance for country i and is the distance between countries i and j for all j not equal to i.

To illustrate, consider the example of New Zealand. First we first take the NZ GDP and divide it by the (physical) internal distance within NZ. Then take the Australian GDP and divide it by the physical distance between Australia and NZ. Then we take Japan’s GDP and divide it by the distance between NZ and Japan and so on. Summing these terms, we get a “remoteness” index for NZ of 2.45. The results for OECD countries are shown in Table A1.

Note: Small numbers mean more remote

The data used are as follows.

Internal distance of a country is calculated as the square root of a country’s total area in miles times 0.4. So, for example, New Zealand’s total area is 169,084 square miles. The country area is taken from The Times of London: Concise Atlas of the World, 8th Edition (2000).

Distance between countries is the Great Circle distance between capital cities (see http://www.wcrl.ars.usda.gov/cec/java/capitals.htm).

GDP is current gross domestic product is that reported by the OECD for 2001. The GDP data in domestic currency units are into U.S. dollars using market exchange rates (see: http://www.wcrl.ars.usda.gov/cec/java/capitals.htm).

The above calculations have not been varied. One would want to do a sensitivity analysis on some of the parameters adopted. Most critically, internal distance and international distances are treated symmetrically here; more likely an international mile is more of a barrier than a domestic one. If we were to recalculate these figures assuming that an international mile is more costly than a domestic mile, the U.S. and Japan would move down toward the not remote end of the scale, Canada and some of the more eastern European countries (eg, Poland, Slovakian Rep, Hungary) would move up toward the more remote end of the scale -- because they rely disproportionately on closeness to foreign markets for their ranking in Table A1. In essence, Table A1 reflects only physical distance, so it is the “base” case in a world where there were no country boundaries.