Agriculture will be the make-or-break issue for the United States, the European Union, and the Group of 20 mainly larger developing countries (G-20) at the World Trade Organization ministerial conference in Hong Kong. On the surface, obstacles to an agreement about agriculture seem insuperable. Two years ago, trade talks at Cancún broke down principally because the G-20 rejected U.S. and EU offers in this area. But a careful examination of the current highly complex agricultural trade regime reveals that prospects for an agreement are not as bleak as they appear. After sorting through and assessing the consequences of the different policy instruments now in use, fears of an impasse look misplaced and the outlines of a successful agreement emerge.

Before the Uruguay Round Agreement on Agriculture (URAA) came into effect on January 1, 1995, international trade in agriculture had remained almost entirely outside of the scope of the General Agreement on Tariffs and Trade (GATT). Over the years, powerful farm lobbies in developed countries had sought and received border protections complemented by price supports and export and output subsidies. Developing countries did not object much. Convinced that development was synonymous with industrialization, they focused on seeking access to the developed countries' markets in industrial products.

Given the highly distorted regime that resulted and the continued influence of farm lobbies in economically powerful countries, the URAA was little more than a step in the right direction toward rationalizing, and then dismantling, the inherited protectionist regime. The agreement was implemented in developed countries from 1995 to 2000 and in developing countries from 1995 to 2005. It required each WTO member to replace all border barriers (tariffs, quotas, and combinations of the two) against the imports of agricultural commodities with an equivalent tariff. The replacement tariff was intended to approximate the protectionist effect of existing tariff and nontariff barriers. (This so-called tariffication process was intended to introduce transparency and to rationalize the protectionist regime.) The resulting tariff was then "bound," meaning that it would henceforth define the maximum legal tariff on the commodity.

The URAA required member countries to cut their bound tariffs by a predetermined percentage according to an agreed-upon timetable. But by taking advantage of the high protection in the base period (1986-88) and flexibility in defining a tariff equivalent, virtually all countries managed to bind the tariffs on their import-sensitive commodities at levels substantially higher than the prevailing official tariff, called the Most Favored Nation (MFN) tariff in GATT/WTO terminology. Anticipating that in many cases even the MFN rates were prohibitively high, the URAA required member countries to guarantee a prespecified minimum (de minimis) market access for each product. Most countries chose to meet this requirement by introducing a quota equal to the de minimis obligation and a third, even lower, tariff on imports within the quota. After the quota is filled, the tariff jumps to the MFN rate.

Doha Round negotiations are over the bound rates. Because the bound rate is often substantially higher than the applied rate (which is either the MFN rate or the lower within-quota rate), no actual liberalization will take place in many commodities unless the Doha negotiations bring deep cuts.


Although the URAA tariff regime is complicated, the domestic-subsidy regime the agreement has spawned is substantially more so. Under current WTO rules, countries are free to employ four categories of subsidies: those in the "green" and "blue" boxes, certain development measures, and the separate de minimis subsidies. Subsidies in the green box have no or minimal distorting effect on production and hence trade. They include measures decoupled from output such as income-support payments, safety-net programs, payments under environmental programs, and agricultural research-and-development subsidies. The blue box contains direct payments under production-limiting programs. They cover payments based on acreage, yield, or number of livestock in a base year. Because countries are allowed to revise the base year over time, subsidies in the blue box may have an effect on current output. Development measures cover direct or indirect assistance aimed at encouraging agricultural and rural development in developing countries. They include investment subsidies generally available to agriculture (e.g., research and development, extension programs, and soil and water conservation) and agricultural input subsidies available to low-income or resource-poor farmers (e.g., fertilizer, water, and electricity). finally, under the de minimis provision, developed countries are allowed to use other subsidies with an aggregate value of up to 5 percent of the total value of domestic agricultural production (developing countries can use them up to 10 percent).

The WTO assigns all subsidies outside of the green and blue boxes and development measures -- such as support prices, direct production subsidies, and input subsidies, including those permitted under the de minimis rules -- to an "amber" box. These are generally trade-distorting and therefore the proper subject for reduction in the multilateral trade negotiations. Hence, the URAA targeted the Aggregate Measurement of Support (AMS), defined as the amber-box subsidies net of de minimis subsidies. It required member countries to report their total AMS for the period between 1986 and 1988, bind it, and reduce it according to an agreed-upon schedule. Those reductions have now been fully implemented, but there remains a large gap between the bound and the applied AMS.


Although export subsidies are very much in the news, they are no longer a major source of trade distortion, at least in aggregate. The URAA required all countries to report, bind, and reduce their export subsidies, as with AMS. Countries reporting no export subsidies on a commodity had to bind the relevant subsidy at zero. Only 25 WTO members, including nine developing countries, reported having any export subsidies, and all of the subsidies applied only to a limited set of commodities. Like tariffs, bound rates for export subsidies are higher than applied rates.

The latest year for which complete data from the WTO on export subsidies are available is 1998. The total amount of export subsidies spent by all WTO members that year was $5.4 billion. The EU accounted for $4.95 billion, followed by Switzerland, with $292 million, and the United States, with $147 million. Because the level of U.S. subsidies dropped to $80 million in 1999 and of EU subsidies to $2.6 billion in 2000, it is safe to assume that total export subsidies are currently probably less than $3 billion worldwide and definitely no more than $5 billion. Surprisingly, therefore, the total elimination of export subsidies would be not a big leap but a small step.

Meanwhile, trade-distorting domestic subsidies are much larger than export subsidies, but the reality is less dire than is commonly assumed. Under the URAA, only 34 countries reported AMS subsidies in the base period. Although many developing countries, such as Argentina, Brazil, Mexico, South Korea, South Africa, and Thailand, were in this group, their subsidies were small relative to those in the major developed countries; subsidies are more difficult to finance in poor countries with tighter budget constraints.

Based on the latest and most complete data available from the WTO, total AMS provided by the top five domestic-subsidy users -- in order, the EU, the United States, Japan, Switzerland, and Norway -- was $71.1 billion in 1998. (The EU's share was $51 billion and that of the United States was $10.4 billion.) Extending the list to the top ten users, thereby also including Mexico, South Korea, Canada, Israel, and Thailand, increases the total to only $74.8 billion. Even including blue-box and de minimis subsidies (permitted under the current WTO rules) yields a total of only $100.7 billion for the top five and $106 billion for the top ten (with Brazil replacing Thailand on the latter list). And all indications from the available information are that these numbers are smaller today than they were in 1998. The EU, in particular, has been moving steadily toward turning its amber-box subsidies into green-box subsidies.


Yet the numbers commonly cited in the debates about subsidies are far larger. A recent New York Times editorial on the Doha Round repeated the often-cited figure that developed countries spend $1 billion per day in agricultural subsidies. World Bank President Paul Wolfowitz referred to developed countries expending "$280 billion on support to agricultural producers" in a recent op-ed in the Financial Times. Oxfam routinely accuses rich countries of giving more than $300 billion annually in subsidies to agribusiness.

The press, nongovernmental organizations (NGOs), and senior staff members at international financial institutions have been either naïve or disingenuous in making these inflated claims. Instead of using estimates of export subsidies and amber-box subsidies (the only subsidies all WTO members consider trade-distorting), they have chosen to focus on an altogether different concept, the Organization for Economic Cooperation and Development's Producer Support Estimates (PSE). It is misleading to call these subsidies: the PSE measures total revenues earned by producers over the world price, whether as a result of tariff protection, export subsidies, output subsidies, or price-support programs. This means that even if there were zero subsidies as conventionally defined but tariffs or quotas raised the domestic price above the world price, the PSE would be positive. Few economists would approve of such a definition of subsidy.

Routine assertions regarding tariffs are also mistaken and misleading. Contrary to complaints by NGOs, media, and international bureaucracies about developed countries' "double standards," developing countries more than match developed-country tariff protection in agriculture. Calculations of trade-weighted average applied tariffs in 2001, the latest year for which data are available, prove the point. At the aggregate level, the rate for such tariffs was 14.3 percent in developed countries and 20.9 percent in developing countries. The rate was 35.5 percent in Japan, 28.6 percent in the European Free Trade Area, 11.8 percent in the EU, and 2.7 percent in the United States. The rate was 93.9 percent in South Korea, 44.1 percent in India, 38.9 percent in China, 30.4 percent in Pakistan, 25.6 percent in those sub-Saharan African states not classed as least-developed countries (LDCs), and 12.9 percent in Mercosur (the South American trading bloc including Argentina, Brazil, Uruguay, and Paraguay).

These aggregate rates mask important variations across commodities. In particular, commodities for which countries had to introduce a quota and a tariff rate lower than their MFN rate under the URAA are subject to disproportionately high tariffs. For example, the trade-weighted average applied tariff on these goods (including, in the United States, certain dairy products, beef, cotton, peanuts, sugar, some products containing sugar, and tobacco) was 36.9 percent in developed and 63.7 percent in developing countries. The rate reached 103 percent in Japan and 226 percent in South Korea. Thus, any liberalization must focus specifically on this set of products.

Sugar is highly protected in virtually all major developed and developing countries. It is subject to the following MFN rates, for example: 72 percent in South Africa, 60 percent in India and Japan, 56 percent in high-income developing Asia, 43 percent in the United States, 23 percent in Central America and the EU (and 74 percent in other European countries), 18 percent in China, and 17 percent in Argentina and Brazil. Thus, reforming tariffs on sugar will require virtually all WTO members to liberalize. The EU and the United States are major offenders, but others -- including developing countries -- are not without blame.

Although bound tariffs are much higher than applied tariffs worldwide, the gap is much larger in developing countries. In developed countries, the average bound rate is nearly twice as large as the applied rate; in developing countries, it is two and a half times larger. Bangladesh offers the most egregious example: its average bound rate is more than ten times the average applied rate. India, Pakistan, and sub-Saharan Africa also have bound rates more than three times as large as their applied rate. Therefore, unless tariff cuts are deep, and the special treatment that privileges developing countries to lower cuts is checked, there will be little reduction in the applied rates in many developing countries.


Such tariffs, export subsidies, and domestic subsidies have two features in common: they raise prices received by producers in the countries providing them and lower prices in the world market. For example, a tariff or export subsidy set by the EU diverts sales from the EU to the world market, thereby raising the internal EU price and lowering the world price. An EU output subsidy increases the output of the product in the EU and lowers the unit price in the EU and in the rest of the world by less than the subsidy per unit, causing the EU producers' prices and revenue to rise.

Producers in the country employing these interventions necessarily benefit because both the unit price they receive and the total quantity they sell rise in all cases. Not surprisingly, they oppose dismantling these instruments. Countries that import the products subject to these interventions also benefit from reduced world prices. Poor countries that enjoy duty-free access to the markets of tariff-levying countries also benefit because they are able to sell their exports at these countries' internal prices. Thus, they have the same protection as the producers of the tariff-levying countries and are winners, not losers, in the current agricultural regime. This is particularly the case with LDCs, which enjoy duty-free access to the EU market under its Everything But Arms initiative. On the other hand, food-exporting countries such as the members of the Cairns Group, which includes the world's most competitive agricultural producers, are hurt by lower world prices and therefore have the greatest incentive to seek liberalization. For the same reason, the overall impact of the interventions, especially of subsidies, is also negative on the EU. In technical jargon, the terms of trade deteriorate for the EU.

These observations yield some key conclusions. The common assertion that agricultural liberalization in rich countries would bring large benefits to LDCs is mistaken. These states -- many of them poor African countries -- benefit from the current regime because they can sell their exports at the high EU prices and buy imports at the low world prices. (Cotton is perhaps the sole exception: U.S. subsidies hurt poor countries because the EU tariff on cotton is zero and therefore its internal price for cotton is the same as the world price.) Gains to those developing countries not in the Cairns Group would accrue principally from their own liberalization. The principle of comparative advantage applies just as much to agriculture as to industry. Moreover, because developing countries do not currently enjoy trade preferences in one another's markets, they stand to gain from access there.

Meanwhile, liberalization in developed countries would principally benefit them. Ending their agricultural subsidies would eliminate not only inefficiencies but also the losses from the spillover of the subsidies to the importing countries. Cutting tariffs will generate benefits for their consumers by lowering prices. And countries with a comparative advantage in agriculture -- mainly developed countries such as the United States, Canada, Australia, and New Zealand as well as the richer developing countries in the Cairns Group such as Brazil, Argentina, Malaysia, and Indonesia -- would benefit from the higher world prices that would follow liberalization in the developed countries.

Gains from the removal of subsidies under the Doha Round, moreover, are likely to be much smaller than previously thought. For one thing, negotiable subsidies have never been as large as has been publicized, and they have declined in importance over the years. Today, export subsidies are in the $3 billion to $5 billion range and domestic subsidies subject to negotiations are well below $100 billion. These numbers are not insignificant, but they are much smaller than commonly believed, making tariffs the more serious barrier to agricultural trade.

Numerous estimates by economists support these conclusions. Careful research shows that the gains to developing countries outside of the Cairns Group from developed-country liberalization would be meager, and that countries that today enjoy preferences in the rich-country markets typically would lose from such reform. Developing countries, especially small ones and those with high initial protection of their own, would benefit more from their own liberalization.


That agriculture remains the most contentious issue on the table is unfortunate. After all, the progress made during the last decade suggests that further liberalization in this sector should be within reach. For instance, the export subsidies that were the center of attention not long ago are now in the $3 billion to $5 billion range. Domestic subsidies subject to WTO discipline are down to well below $100 billion. Some may complain that the United States and the EU have inappropriately shifted many trade-distorting subsidies into the green or blue box. But even if this is so, the adversely affected member countries can challenge such moves through the WTO dispute-settlement body, as Brazil recently did successfully with some U.S. cotton subsidies.

What could be a realistic way to conclude the Doha Round? Export subsidies are perhaps the most straightforward issue to tackle. Given their insignificant level today, eliminating them by 2010, as proposed in the latest U.S. offer, is possible. Admittedly, the economic benefits of such a reform are likely to be small, but it would give participants in the negotiations a significant psychological boost by ending an entire category of subsidies.

As for domestic subsidies, the first point to emphasize is that there remains a significant gap between bound and applied amber-box subsidies. Based on the latest data available, this gap was approximately 35 percent for the EU in 2000 and 25 percent for the United States in 2001. It is safe to assume that these gaps are larger today. Given this fact, the current U.S. proposal that the United States reduce these subsidies by 60 percent and the EU reduce them by 80 percent is achievable: in effect, it is less than meets the eye, because the reductions in the applied rates would be substantially smaller.

A different issue pertaining to domestic subsidies concerns the de minimis and blue-box subsidies. The aim now must be to end both these measures to achieve a clean subsidy regime. De minimis subsidies are clearly trade-distorting and should be phased out. As regards the blue box, the objective of the reform should be to simply eliminate the category, moving the subsidies that distort trade to the amber box and those that do not to the green box. These reforms will increase transparency by simplifying the subsidy structure, as has been done with subsidies on manufactures.

Ideally, the regime would contain just two categories of subsidies. One category -- the new green box -- would contain measures such as income-support payments and agricultural-research subsidies that do not distort trade. These would be unrestricted, because the provision of subsidies that do not affect production and trade should be an internal matter. The second category, the red box, would contain measures that do distort trade, such as output subsidies and price supports, and therefore would be prohibited.

The largest potential gains for all parties involved, especially developing countries, however, would come from lowering border barriers, mainly tariffs. In the August 2004 Framework Agreement setting the detailed terms of the Doha negotiations, member countries accepted the principle that high tariffs should be subject to deeper cuts than low tariffs. This makes sense for two reasons. first, because higher tariffs have a larger gap between the bound and applied rates, higher cuts will be needed to achieve cuts in the applied rates. Second, cutting higher tariffs more and lower tariffs less will reduce the large variations in tariff rates across commodities.

The current disagreement surrounds the extent of the cuts. The United States proposes that the top tier of bound tariffs be cut by 90 percent, with average cuts of 75 percent. The EU has suggested a less ambitious approach that cuts the highest bound tariffs by 60 percent, with average cuts of 46 percent. (The G-20 is arguing for average tariff cuts of 54 percent.) The EU has also put forward a proposal to allow countries to designate up to 8 percent of imports as "sensitive" and exempt them from the bulk of the cuts. The United States proposes to limit such exceptions to just 1 percent of the products.

U.S. insistence on larger tariff reductions and fewer exceptions is justifiable for at least two reasons. First, shallow cuts in bound rates would leave the actual tariffs unchanged, especially for high tariff rates. A long list of sensitive products would likewise allow countries to exclude entirely the products with highest tariffs and therefore forego the largest potential trade gains from liberalization. Second, Washington believes that by exempting LDCs from reciprocity obligations under the Framework Agreement, it has already made a major concession and needs reciprocity on agricultural tariffs from the EU, Japan, and more advanced developing countries if it is to persuade its own farmers to give up their subsidies and protection.

On the other hand, the EU also needs compensation for its concessions. Recall that at Cancún it dropped investment, competition policy, and government procurement from the Doha agenda. And because the EU does not have a comparative advantage in agriculture, it is naturally seeking cross-sector reciprocity in the form of liberalization in industrial products and services. The next step in breaking the U.S.-EU impasse is to put offers on industrial products and services on the table quickly. This would be a step forward: the elimination of tariff peaks in developed countries and liberalization by developing countries in trade in the industrial sector promise gains commensurate with agricultural liberalization.

But for tariff reductions to really be beneficial, action will be required of both developed and developing countries. The gains to developing countries from lowering border barriers will be minuscule if reform is limited to developed countries. Unfortunately, the Framework Agreement has already done a disservice to LDCs by exempting them from liberalizing agriculture. Their own liberalization could have at least partially counteracted the adverse effects of diminished access to developed-country markets caused by preference erosion and rising health and safety standards on their exports, a handicap the bigger and more developed Cairns Group exporters, which are better able to handle such protectionist tactics, do not face to the same degree. LDCs wishing to minimize the damage from this well-intentioned but misguided exemption would be well advised to liberalize unilaterally and seek as much technical assistance as possible from developed countries and international financial institutions to meet the health-and-safety standards importing countries impose. The Framework Agreement also allows developing countries to declare certain products "special" and apply a safeguard clause to prevent import surges. If the list of special products is too long and the safeguard too easy to invoke, any liberalization achieved in developing countries will be undermined.

Finally, aid from multilateral institutions such as the World Bank will be necessary to make liberalization attractive to developing countries. Countries that experience preference erosion will need to be compensated, at least in the short run. Estimates of these losses vary widely, but a figure in the neighborhood of $1 billion cannot be ruled out because the losses to producers in the sugar trade alone are estimated at $450 million. Aid is also required to assist countries in meeting the adjustment costs of liberalization. Developed countries can afford trade-adjustment programs. Developing countries cannot. Given the importance the World Bank attaches to trade liberalization, it should promote an "aid-for-trade" program.

Contrary to the impressions conveyed by the media, definite progress has been made toward rationalizing agricultural policy and lowering trade-distorting subsidies over the past ten years. Moreover, the differences among the major players are hardly insurmountable. Some flexibility will be required to produce an outcome that will benefit everyone. Negotiators must not forget that, unlike the Uruguay Round, which dealt with contentious issues such as intellectual-property protection, this round is focused on trade liberalization. And even if they must make other concessions, liberalizing trade will benefit all countries.

You are reading a free article.

Subscribe to Foreign Affairs to get unlimited access.

  • Paywall-free reading of new articles and a century of archives
  • Unlock access to iOS/Android apps to save editions for offline reading
  • Six issues a year in print, online, and audio editions
Subscribe Now
  • ARVIND PANAGARIYA is Professor of Economics and Jagdish Bhagwati Professor of Indian Political Economy at Columbia University.
  • More By Arvind Panagariya