The State of Agricultural Commodity Markets 2022


Comparative advantage, trade policies and trade costs

In 2019, global value added in food and agriculture amounted to USD 3.7 trillion, four times less than that in manufacturing (USD 13.7 trillion). For the same year, the value of food and agricultural exports was USD 1.5 trillion, eight times less than that of manufacturing (USD 12.7 trillion). This significant difference between the value added and trade in these sectors implies that food and agricultural products are not traded as intensively as manufactures.

Another interesting point is that in low-income countries, the share of food and agricultural imports in consumption is low relative to food that is produced domestically (see Figure 2.10). There are many reasons that explain why some food products are produced and consumed locally, but as low-income countries are, in general, characterized by low agricultural productivity relative to high-income countries, one would expect that they would be relatively larger importers of food.

Both of the above observations have led analysts to refer to the “puzzle” of missing trade in food and agriculture.o Looking at this puzzle offers an insight into how comparative advantage determines trade flows in food and agriculture, as well as into the role of trade policies and trade costs in lessening comparative advantage’s strength.

Productivity and comparative advantage

Across countries, productivity differences in agriculture are significant compared to other sectors of the economy. Figure 2.1 suggests that agricultural productivity per worker – measured as value added per worker – is much lower than in non-agriculture for most countries (as most observations lie above the diagonal line). The data also suggest that, across countries, the heterogeneity of productivities per worker in agriculture is much higher than in the non-agriculture sectors (Table 2.1).p

FIGURE 2.1Productivity in agriculture and non-agriculture sectors, 2019

SOURCE: World Development Indicators, World Bank Group.
NOTE: Relationship between value added per worker in agriculture, forestry and fishing and non-agriculture, that includes manufacturing, mining and quarrying, construction and utilities, across countries.
SOURCE: World Development Indicators, World Bank Group.


SOURCE: World Development Indicators, World Bank Group.
NOTE: Non-agriculture includes manufacturing, mining and quarrying, construction and utilities.
SOURCE: World Development Indicators, World Bank Group.

The top 10 percent of the richest countries produce 70.4 times as much agricultural value added per worker as countries in the bottom 10 percent of the income distribution. For example, in the United States of America, agricultural value added per worker in 2019 amounted to USD 100 062 (measured in 2015 prices) as compared with an average of USD 944 in sub-Saharan African countries, including Burkina Faso, the Democratic Republic of Congo, Ethiopia, the Gambia, the Niger and Mozambique.

Productivity differences in the non-agriculture sector are also pronounced but their heterogeneity is significantly smaller relative to agriculture (Table 2.1). On average, countries in the top 10 percent of the world income distribution produce 40.2 times as much non-agriculture value added per worker as the bottom 10 percent countries.

Across countries, the heterogeneity of productivities per worker in agriculture is much greater than in non-agriculture, reflecting a potentially powerful influence of comparative advantage on trade flows. Nevertheless, food and agricultural products are not traded as intensively as non-agricultural ones.

Agriculture is unique in that, on average, technology accounts for about three-quarters of productivity growth at the global level, while increases in the factors of production, such as land, make up for one-quarter of productivity growth.89 However, the agricultural productivity gap between high- and low-income countries is vast (see Figure 2.2) and these cross-country productivity differences in agriculture have been the subject of much research. Researchers suggest that the large productivity gap in agriculture can be due to significant barriers to technology adoption and limited access to modern inputs.90

FIGURE 2.2Agricultural productivity and Gross Domestic Product per capita, 2019

SOURCE: World Development Indicators, World Bank Group.
NOTE: Relationship between value added per worker in agriculture, forestry and fishing and gross domestic product per capita across countries.
SOURCE: World Development Indicators, World Bank Group.

In low-income countries, market failures can inhibit technology adoption. High transaction costs result in thin markets and as farmers’ participation in markets is low, they face only localized demand, which becomes quickly satisfied with small increases in production.91 This provides weak incentives to farmers to adopt new technologies and increase productivity.

Uncertainty also affects the decision to adopt new technologies. Smallholder farmers, for whom the additional cost of modern technology would make up for a significant share of their income, are risk-averse and prefer to use traditional technologies.92 Farm size plays an important role in shaping farmers’ attitudes to risk and technology adoption, and the differences in farm size between high- and low-income countries could also explain a large part of the agricultural productivity gap.93, 94 Incomplete insurance markets also result in low rates of technology adoption. Farmers in developing countries tend to use fewer modern inputs, such as fertilizers, because of uninsurable risks.95

Women farmers face even greater disadvantages than their male counterparts, as they have less access to knowledge and social capital, which are additional factors that determine productivity.96

Economy-wide factors also contribute to low agricultural productivity per worker in low-income countries. Poorly functioning labour markets together with a lack of education and low skills in rural areas can inhibit the reallocation of labour from agriculture to other sectors of the economy, thus contributing to the agricultural productivity gap.97, 98 Fewer agricultural workers would translate into additional gains in productivity per worker, but for this to happen labour markets should function well.

Trade policies

Domestic support and trade policy measures in food and agriculture address a broad array of objectives. For example, domestic support, such as input subsidies, aim to improve farmers’ access to inputs. Direct income support measures contribute toward maintaining a level of farm income that keeps pace with the income trends in other economic sectors. Tariffs can be used to protect local farmers from international competition, reduce import dependence and promote self-sufficiency in staple foods. Export restrictions can lower the domestic price of food and contribute towards food security in the short term. Both tariffs and export taxes provide an important source of government revenue. Such policies can distort prices and influence trade.

NTMs are effectively trade policy measures in the sense that they can have an economic effect on trade, changing quantities traded or prices or both. NTMs include sanitary and phytosanitary (SPS) measures that ensure food safety and protect animal or plant health, as well as other technical regulations and standards, referred to as technical barriers to trade (TBT), that relate to objectives such as environmental protection, labour health and safety, and prevention of deceptive practices (see, for example, the discussion on environmental measures in Part 3). The relationship between NTMs and trade is complex. Many NTMs may restrict trade but address important issues that improve welfare. At the same time, they can also expand trade as they strengthen the demand for a product through better information on its health and sanitary characteristics.

In the context of trade policy literature, both tariffs and NTMs contribute to trade costs and could partly offset the influence of comparative advantage on trade flows between countries (see Box 2.3 for a definition of trade costs and a discussion on their measurement).

BOX 2.3Trade costs and how they are measured

In the context of the trade literature, all factors that drive a wedge between prices in exporting and importing countries give rise to trade costs and influence trade flows. This definition of trade costs includes trade policies, such as tariffs and NTMs. Although tariffs, both ad valorem and specific, are directly observed and their impact on trade flows can be assessed relatively easily, the costs and trade effects of other trade policies, such as NTMs, are difficult to observe. For example, the application of a maximum residue level for pesticides to imports may increase or decrease trade or could result in a rejection of shipments, depending on whether imports comply with the regulation. Other trade costs, such as transport costs, administration and transaction costs, and costs arising due to border delays are also inherently difficult to observe, or the data available is not adequate to support measurement. Distance, common language, information availability and regulation enforcement also play a role in determining trade costs.

Observable costs, such as freight rates and tariffs, can be assessed without difficulty but in order to measure costs that relate to informational and institutional factors and NTMs, analysts turn to economic models. These models link trade flows to observable variables, such as price differentials, common language or common borders, distance or participation in a trade agreement, and they take into account the unobservable costs by linking trade flows to their theoretically predicted values.137, 138 Often, these modelling approaches capture a wide range of trade costs, including tariffs, as an ad valorem equivalent.

An important initiative by the Economic and Social Commission for Asia and the Pacific (ESCAP) and the World Bank Group uses modelling frameworks to estimate trade costs. The analysis in this report also uses price data and modelling to assess trade costs in food and agriculture. On the basis of these models, analysts conclude that trade costs are high and play an important role in shaping trade. For example, the assessment suggests that trade cost declines explain roughly 33 percent of the post-World War II trade boom.139


With the Uruguay Round agreements, including the AoA in 1995, members of the WTO committed to not restricting imports by any means other than tariffs, and to keep their rates within set thresholds determined for each country. Many countries apply lower tariffs than the maximum permitted level. This unilateral reduction in tariffs, instigated by the AoA, together with concessions made in regional agreements, resulted in a substantial liberalization of trade.

The reduction in applied tariffs has been significant. Multilateral, unilateral and regional concessions are estimated to have contributed to a reduction of about 27 percent in average food and agricultural tariff levels worldwide. These reductions brought greater market openness and promoted trade significantly (see Part 1).

Nevertheless, the extent of tariff reduction in low- and middle-income countries was less than in high-income economies (Figure 2.3). Analysts suggest that this process of reduction in applied tariffs by low- and middle-income countries became less significant after the 2008 financial crisis.99

FIGURE 2.3Applied tariffs in agriculture, 1995–1999 and 2016–2020

SOURCE: UNCTAD-TRAINS data from World Integrated Trade Solution.
NOTE: Average of effectively applied tariffs.
SOURCE: UNCTAD-TRAINS data from World Integrated Trade Solution.

The process of lowering tariffs was more effective in non-agriculture sectors. On average, tariffs applied on industrial goods are significantly lower than on agriculture (Figure 2.4). Many low- and middle-income countries lowered trade barriers for manufactures and other industrial products to promote participation in global value chains. Agricultural tariffs remain relatively higher, especially in low- and middle-income countries, implying a relatively higher rate of protection for the sector and, a potentially larger negative impact on the influence of comparative advantage (see Box 4.1 on the political economy of protection of food and agriculture).

FIGURE 2.4Applied tariffs in agriculture and manufacturing, average 2016–2020

SOURCE: UNCTAD-TRAINS data from World Integrated Trade Solution.
NOTE: Average of effectively applied tariffs.
SOURCE: UNCTAD-TRAINS data from World Integrated Trade Solution.

Non-tariff measures

NTMs are more prevalent in agriculture compared with other sectors and this contributes to relatively higher trade costs in agriculture.100 Food and agricultural trade is subject to the highest incidence of NTMs, both at the intensive and at the extensive margin – that is, in terms of total trade value and the number of products traded respectively. Almost 100 percent of food and agricultural imports are subject to NTMs compared to an average of 40 percent for all other sectors. Food and agricultural products are heavily regulated and subject to the highest number of NTMs per product. On average, a food product faces eight different NTMs compared to just under two NTMs for products of all other sectors.101

NTMs increase the cost of trade, particularly if the importing country applies different regulations than those applied by the exporter. In this case, exporters face additional trade costs related to: identifying and processing information on the relevant requirements in the import markets; adjusting the production process to these requirements; and proving that these requirements are met.102 Recent evidence from the analysis of regulations in 110 countries suggests that the trade costs associated with NTMs can increase import prices of agricultural products by nearly 15 percent in ad valorem equivalent.103 There are also implicit costs that are associated with NTMs. Firms that export to different destination markets and face different standards, as for example, different labelling requirements, must produce different versions of their products, which incurs significant costs in terms of efficiency and foregone economies of scale.104, 105

Low-income countries face much higher NTM-related costs than high-income countries. NTMs are widespread in food and agriculture and food and agricultural products make up a significant part of exports by low-income countries. Poor transport and communication infrastructure, low organizational and technical capacities make the cost of compliance with standards higher in low-income countries than in developed economies. It is estimated that in low-income countries, NTMs result in an additional 3 percentage points in ad valorem trade cost equivalent compared with high-income countries.106

While SPS and TBT provisions raise trade costs, they can also enhance trade by strengthening the demand for imported products. In food and agriculture, compliance with SPS measures – essential for ensuring the health of consumers, animals and plants and the protection of the environment – increases consumer confidence in imported products.107 Harmonizing NTMs across countries is important to reduce their costs and to enhance trade. Often, RTAs include provisions for deeper cooperation in regulation and standards to promote trade among their members (see Parts 3 and 4).q

Other trade costs

A large body of analytical work focuses on the costs that capture frictions in trade. In addition to the costs related to the NTMs, these include direct costs, such as freight and insurance costs, and indirect costs, such as costs related to export and import procedures, legal and regulatory fees, expenses associated with the use of different currencies and different languages and time delays at the border.108 These trade costs are rarely measured directly, as they are not observable as tariffs, but are typically estimated and inferred from models (see Box 2.3).r

Despite the focus on globalization, a significant part of trade takes place between countries that are geographically close to each other (see Part 1). Physical distance increases trade costs and makes trade with neighbouring countries relatively cheaper. Empirical research on a wide range of distance effects suggests that, on average, a 10 percent increase in distance results in a decrease of about 0.9 percent in trade flows.109 Distance matters, and its role in determining trade costs and trade flows is significant. Although this negative relationship between distance and trade is persistent, trade costs vary significantly across both goods and countries.

Trade costs tend to be much higher for food and agriculture than for other products, such as manufactures. For example, bilateral trade flows between Kenya and Uganda – low- and middle- income countries that share borders in sub-Saharan Africa – are subject to an ad valorem trade cost equivalent of 130 percent for agricultural products and of 78 percent for manufactures.s Differences in trade costs between agricultural products and manufactures are also observed in high-income countries. For example, the ad valorem trade cost equivalent in agricultural products between France and Germany – two neighbouring members of the European Union – amounts to 65 percent, while the corresponding value for manufactures is 31 percent.

Such high trade costs can inhibit international trade, and, unsurprisingly, trade intensities between food and agriculture and manufactures differ. Low value-to-weight ratios of food and agricultural products result in higher trade costs than manufactures. It takes much more fuel and stowage to move an amount of wheat worth USD 1 000 than shipping USD 1 000 of mobile phones and, therefore, the increase in the import price of wheat due to freight costs relative to that of mobile phones is much higher.

There are also implicit costs that are difficult to measure. For perishable agricultural products, border delays can be particularly costly. On average, estimates suggest that for food and beverages, a delay at the border of one day is equivalent to a 3.1 percent ad valorem trade cost, compared with 2 percent for consumer and capital goods.110 Another study finds that delay-related import costs for agriculture in low-income countries could amount to up to 400 percent ad valorem equivalent as compared with 30 percent for high-income countries.111

In general, poor infrastructure, weak institutions and market failures result in high trade costs in many low-income countries. For example, a poorly developed transport network results in high transport costs, while low administration capacity and information asymmetries can give rise to significant costs related to delays at the border. The difference in trade costs between low-income and high-income countries can be significant. Trade costs tend to be not only higher for food and agriculture relative to other sectors, but low-income countries face even higher agricultural trade costs than high-income countries.

back to top TOP