No matter which statistics we look at, real growth in trade or trade to GDP ratios, there is little doubt that over the last 20-25 years we’ve seen unprecedented growth in global trade flows. The only exception was the period of the global financial crisis when global trade plummeted but even then recovery was very swift and global trade bounced back. In 2016, global trade accounted for about 60 per cent of world GDP which is 20 percentage points higher than at the beginning of the nineties.

The growth in trade is, of course, linked to trade liberalisation measures taken by governments around the globe. For example, at the beginning of nineties about 50 regional trade agreements were in force, whereas there are currently about 270 enforced agreements. At the same time, it is clear that somewhere along the way, many people’s attitude towards trade liberalisation and the free movement of goods and labour drastically changed. The backlash against globalisation is a reality in many countries. The UK voting for Brexit and the US pulling out of the Trans-Pacific Partnership are merely the two most striking examples of this change in attitudes. This growing skepticism towards globalisation can be explained, in part, by the unequal gains from trade across different population groups.

People differ in how they spend their income. In Figure 1, I plot the share of income spent on food across the different deciles, d, of a given country’s population and sorted from the poorest to the richest decile along with the aggregate country-specific expenditures. It is striking that within-country differences in spending patterns are more acute than aggregate differences across countries and this must be taken into account when calculating the welfare gains from trade. In other words, aggregate gains from trade may not be informative about the individual-specific effects of trade on real income and thus should be considered with caution.

Figure 1: Share of income spent on food

Source: Nigai (2016)

However, it has become standard in the trade literature to talk about the effects of free trade using the notion of the average consumer. These results are very useful when talking about aggregate gains from trade but provide only a vague idea of what happens to different groups of people upon trade liberalisation. This simplifying assumption that we have employed for many years has created a disconnect between economists and the general public. The reason for this is very simple, people differ in many ways and some are just too different from this elusive average. Therefore, in reality, it may be challenging for some groups to relate to the aggregate gains from trade.

Consider the following example. Let’s group goods into three large categories: agricultural goods that cover food items, manufacturing goods that cover things like clothing and electronics, and services that cover activities such as going to restaurants. Let’s imagine two groups of consumers with different levels of income. The low-income group will spend a large share of their income on food and less on manufacturing goods and services. At the same time, the high-income group will do quite the opposite: the majority of their income will be devoted to manufacturing goods and services and very little to food items. Now suppose trade liberalises and prices change for food, manufacturing goods and services. Who will benefit more from the liberalisation?

The answer to the aforementioned question will depend on which good will experience the highest decrease in prices upon trade liberalisation. In Nigai (2016), I find that the prices of manufacturing goods fall relatively more in comparison to agricultural goods in response to an equivalent reduction in their respective trade barriers. This is due to the finding that the technology of producing manufacturing goods such as cars and computers is more dispersed than the technology of producing agricultural goods such as potatoes and apples. International trade allows people to buy goods from other countries, produced with potentially more cost-efficient technology, and thus at lower prices. Because technology is more dispersed in manufacturing goods than in agricultural goods the prices of cars and computers fall more than those of apples or potatoes.

Because high-income people spend a larger share of their income, in relative terms, on manufactures and services, they will benefit more from trade. In contrast to that, low-income individuals often spend a large share of their income on food and as the price of food items decrease to a much lesser extent, those individuals will gain relatively less.

Importantly, the gains of the average person will reflect neither the larger gains of the rich nor the smaller gains of the poor. In a scenario with a 15 per cent reduction in non-tariff trade barriers, the gains of the rich would be up to 5 percentage points higher than the gains of the average consumer. At the same time, the welfare gains of the poor would be up to 11 percentage points lower than the gains of the average person. This makes the gains of the average consumer uninformative.

There are two main takeaways from this column. First, it is necessary to recognise that the effects of trade liberalisations such as the Transatlantic Trade and Investment Partnership or trade restrictions following Brexit will vary tremendously across different people. So when we inform the policy debates we have to take these differences into account in evaluating how free trade affects people along the income distribution. Second, while potentially everyone can gain from free trade, the gains can be highly unequal with the rich benefitting relatively more. From that perspective, if trade is used as an instrument to help the poor we should target trade facilitation in goods that are essential for them.

Royal Economic Society video in which Sergey Nigai discusses his findings:



Sergey Nigai is an Assistant Professor of Economics at the University of Colorado Boulder. His research focuses on international trade and consumer welfare. His published research papers and working papers can be found at .