Many poor and developing countries get exemptions or partial exemptions from standard developed country tariff rates. These exemptions, or "tariff preferences" are meant to give them a competitive edge, and to help them grow.
Trade negotiations that reduce standard developed country tariff rates erode the effective size of these preferences, and the competitive advantages they provide ("preference erosion"). This could happen even if the preferences were left untouched.
These preferences may or may not be a good thing, economists and others are divided. But they certainly create a practical problem for the WTO's current Doha Round of trade negotiations. The WTO reaches decisions about reducing trade barriers on the basis of consensus among member countries. Many of these countries receive preferences whose value could be eroded by tariff reductions. Since they have a stake in high standard tariff rates, will they ever agree to a substantial liberalizing reform? If not, how will the WTO's members reach a consensus?
So it's worthwhile taking a closer look at these preferences.
Two recent IMF studies suggest that the preference erosion issue is important to a relatively small number of countries, and for a smaller number of products.
One study looked at the poorest countries, and the other looked at middle income countries. Both studies look at potential impacts of changes in the tariff preferences provided by the "Quad" countries: the U.S., Europe, Canada, and Japan. The studies used somewhat different methods, but both focus on the impact of a hypothesized 40% preference erosion in the Quad markets. The 40% is somewhat above the impact from the last major round of trade negotiations. Both studies use methods that tend to provide upper bound estimates of the impact on the countries that get the preferences.
Poor countries
Arvind Subramanian looked at the impact of tariff preference erosion on the world's poorest countries (less developed countries, or LDCs) in the 2003 IMF report, "Financing of Losses From Preference Erosion." (To get this, use the WTO document search page, and type WT/TF/COH/14 - just like that - in the "document symbol" box)
Estimates of the losses from preference erosion for these countries are pretty small. He finds that 40% reductions in the standard manufacturing and agricultural tariffs of the Quad nations, would lead to a 1.7% reduction in the aggregate value of LDC exports. A low average could still conceal large impacts on individual countries. In this case:
"For 26 out of the 46 LDCs, export losses from preference erosion are estimated to be less than two per cent, and for another 15 countries less than five percent. Thus, only five countries may face losses exceeding five percent of exports. These are Malawi (11.5 per cent), Mauritania (8.8 per cent), Haiti (6.4 per cent), Cape Verde (6.3 per cent) and Sao Tome and Principe (5.2 per cent). In absolute value, the largest likely losers are Bangladesh (US$222 million), Cambodia (US$54 million), Malawi (US$49 million), Mauritania (US$40.4 million) and Tanzania (US$29 million). Thus, only Malawi and Mauritania could face a loss that is large in absolute and relative terms."
In a survey paper this spring (see below), Alexandraki comments that Subramanian found that preference erosion would only be significant for a few products - tobacco, textiles, fisheries and cocoa.
Middle income countries
Katerina Alexandraki and Hans Peter Lankes looked at middle income countries. The study itself is "The Impact of Preference Erosion on Middle-Income Developing Countries." (IMF Working Paper WP/04/169. September 2004).
A summary of the results by Alexandraki, with background on the preference erosion issue, and policy implications, just came out in the March 2005 issue of the IMF's magazine, Finance & Development: "Preference Erosion: Cause for Alarm?".
Middle income countries had per capita gross national income between $766 and $9,385 (although the group examined excluded major oil exporters, and European Union accession countries).
Overall, a big reduction in the tariff preference had a small impact on middle income countries:
"...results for the hypothetical case where trade liberalization in the Quad causes a 40 percent reduction in each beneficiary's aggregate preference margin. The impact of preference erosion is found to be small overall (between 0.5 and 1.2 percent of total exports of the middle-income countries considered, depending on elasticities of export supply)..."
(and the authors feel that these estimates are biased upward)
Impacts would be concentrated on a few middle income countries. Of the 76 countries examined, 13 could see their total exports of goods drop by more than five per cent (under the same export elasticity assumption used by Subramanian). Mauritius takes the biggest hit - a loss equal to 19.6 per cent of its goods exports. Other countries taking hits over five percent: St. Lucia, Belize, St. Kitts and Nevis, Guyana, Fiji, Dominica, Seychelles, Jamaica, St. Vincent and the Grenadines, Albania, Swaziland, and Serbia and Montenegro.
And the problems were heavily concentrated on two commodities - sugar and bananas. "Vulnerability to preference erosion is determined overwhelmingly by a country's export dependence on three products, namely sugar, bananas, and - to a far lesser extent - textiles..." and "Sugar and bananas alone explain three-quarters of the current preference margin of countries that have a total margin greater than 5 percent of the value of their exports."
So what?
Too much weight shouldn't be put on the exact percentages; these results are meant to be suggestive, rather than precise.
Alexandraki and Lankes note,
"The discussion of preference erosion, for example in the context of theWorld Trade Organization (WTO), has rarely relied on numerical estimates of the size of the problem. A majority of WTO members benefits from preferential market access in the Quad. This has allowed the impression to prevail that vulnerability to preference erosion is a significant concern shared by a broad range of countries.
The results of this study suggest that the problem is in fact heavily concentrated in a sub-set of products and preference beneficiaries. Reforms in Quad preference regimes for just two products - sugar and bananas - may leave only a fraction of the current preference margins of some of the most vulnerable preference beneficiaries. Accordingly, assistance to help countries cope with preference erosion can be closely targeted at the countries at risk. Many of these are small island economies that may have serious difficulties to adjust."
Revised April 28: This post originally reported the estimate of the loss from preference erosion for Bangladesh as $22 million. This is an error; the actual number is $222 million. The error has been corrected in the post. My thanks to Mr. Wasel Bin Shadat, of the Centre for Policy Dialogue Bangladesh, for pointing this out.