Does the WTO really boost trade

Does the WTO really boost trade?

Wednesday, November 13, 2002


William Watson
National Post

I wasn't in Florence over the weekend, so I don't know how much time the European Social Forum spent denouncing the World Trade Organization but I'll bet it was a lot. According to the left-think that dominated the Florence meetings, the WTO is the corporate oligarchy's main front for propagating the "neoliberal agenda" of a world made safe for capital and dangerous for workers, women, socialists, orphans and all other good people.

If that was indeed the line taken in Florence, the members of the Social Forum need to have a look at a new pair of working papers in which University of California (Berkeley) economist Andrew Rose, a former Torontonian, argues that there is, in fact, no good statistical evidence that the WTO or its predecessor the GATT has either increased trade or liberalized trade policy on a consistent basis across countries or over the years.

The genius of Rose's two papers is simply in asking the questions they do. Most of us, left and right, Florentines and trade fascists alike, simply take for granted that the WTO does create a more liberal world trading regime. What everyone knows sometimes isn't true, however. Credit Rose with the imagination to ask whether any such effect can be seen in the data.

And there is lots of data. Given the large number of countries that now exist and the number of decades over which international agencies have been collecting numbers, Rose was able to look at 200,000 pairs of trading partners. Controlling for other potential determinants of trade, such as the shipping distance between countries, the size of their economies, whether they share a language or have ever been joined in a colonial relationship, and so on, he asked whether the fact that either or both belonged to the GATT/WTO affected the size of the trading flow between them. He found that it did: Trade was actually a little less likely if GATT/WTO member countries were involved. On the other hand, the effect wasn't statistically significant, that is, not big enough not to be a fluke in the numbers. (Toss a coin 1,000 times and chances are it won't come up exactly 500 heads and 500 tails: There will always be a little noise in the numbers.)

Because this is such a counterintuitive result -- almost everybody believes the WTO boosts trade, whether they like that or not -- Rose tested it every which way, adding different explanatory variables to the mix, changing the time periods he studied and divvying up the countries according to continent, income level and so on. There does appear to have been a positive "GATT effect" on trade in the early years and among industrialized countries, but in the great majority of cases membership in the GATT/WTO does not affect the volume of trade in a statistically significant way.

Rose's second paper asks whether GATT/WTO members don't at least have more liberal trade policies than non-members, even if still to be explained forces may have nullified the effect of these more liberal policies on trade flows. One obvious problem here is trying summarize a country's trade policy in a single statistic. Should it be the average tariff rate? Tariff revenues as a share of trade values? The proportion of GDP affected by non-trade barriers? Some composite measure taking into account these and other considerations?

Not wanting to choose a best measure of trade policy, Rose looked at no fewer than 68 separate indexes that have appeared in the economics literature in recent years. And guess what? He found that GATT/WTO members were not consistently more liberal in their trade policies than non-members, though they did consistently score better on the Heritage Foundation/Fraser Institute's index of economic freedom. By most other measures of trade policy, however, member countries were no more nor less liberal than non-member countries.

Do countries not at least become more liberal after joining GATT/WTO? Not necessarily. For instance, in 1986, when Mexico joined the GATT, its average tariff was 6.4% of imports. Five years later, its average tariff was 7.1%. Only after NAFTA did Mexico's tariffs come down sharply. That's consistent with another of Rose's findings, that while membership in GATT/WTO doesn't seem to affect trade, membership in a regional free trade area doubles it -- except in Europe, where formal economic integration has had no statistically significant effect on trade. If it holds up, this finding may tilt the debate about whether regional free trade areas threaten the multilateral process. If FTAs are the only game in town, it may not matter if they do.

Rose himself concedes that his results contradict both common sense and conventional wisdom. And large-scale statistical studies are notoriously prone to all manner of logical pitfall, a fact of life Rose is fully aware of. But you'd think an economic relationship that has led to tear-gassing, busted heads and ransacked McDonald's would shine through in the data. That it doesn't is an intriguing new mystery.