This numerical example of trade creation and diversion is taken from Schiff and Winters (2003, p. 34-36). They suppose that a country can import a good from a potential partner country at €105 per unit and from the rest of the world at €100 and that the same tariff of €10 is levied on imports from both sources, making the prices paid by consumers €115 and €110 respectively. In this situation, consumers will obviously purchase the good from the rest of the world (ROW) supplier and pay €110.
If the country now joins an RTA with a partner, the partners imports come in duty free. The price consumers pay for imports from the partner country then falls to €105, while imports from the rest of the world still cost €110. We expect consumers to switch to the now cheaper source of supply which is the partner country, buying the €105 good and saving €5.
However, because the government now loses €10 per unit (the tariff revenue it was receiving on each unit of ROW imports), the net effect for the country is a loss of €5, and the RTA has reduced real income. The same point can be made another way by pointing out that the country (not the consumer) was previously paying €100 per imported unit and now pays €105. This is the adverse impact of trade diversion.
This example focuses on the importing country, but at the same time that the home country is losing through trade diversion, the partner country is increasing its exports. Can we not argue that the formers loss is just balanced by the latters gain, so that, allowing for some trade creation as a result of the RTA, the region as a whole is better off? Schiff and Winters show that, although there may be a gain to the exporting country, this is less, per unit, than the importer country loss.
Consider the gain to the exporting country when its export sales expand as a result of substituting for ROW exports in the home country market. If exports are just selling at cost (€105 in the example), selling more units does not raise income in the partner country (because some other sector must contract to release resources for the expansion). Thus we are left with a loss in the importing country and no gain in the exporter, resulting in a net loss for the RTA as whole.
If the partner country can sell its exports above cost, then there will be a real income gain. But the highest price the partner exporter can charge will be €110 (above this price, consumers in the home country will switch back to importing from the ROW because, even paying the tariff, ROW imports would now be cheaper). So the maximum income gain to the exporting country as a result of trade diversion is €5 per unit. However, in this situation, there is no gain to consumers in the importing country (they are indifferent as to whether they pay €110 for the good from the ROW or from the partner country) but the government in the importing country is still the poorer by €10 per unit because of the loss of tariff revenue. So overall, the RTA is still worse off by €5 per unit as a result of trade diversion.
This calculation shows that it is possible that one or more partners in an RTA will be a net gainer, but that the RTA will lose overall from trade diversion. For a formal diagrammatic analysis of this proposition, see the Appendix to Chapter 2 in Schiff and Winters (2003).