VoxEU Column International trade

The importance of exports and services trade re-evaluated

In a world economy dominated by fragmented supply chains and trade in tasks, the direct contribution of exports to any national economy is overstated by gross-value measures. Since most measures do not properly account for imported inputs and severely underestimate the share of services in total exports, our view of world trade is distorted. This column says our trade policies risk being distorted too.

It took some time before the concept of vertical specialisation, coined by Hummels, Ishii, and Yi in 2001, reached a broader non-academic audience. Estimating the import content of exports was not something that bothered policymakers. Today however, this formerly academic exercise is at the core of the debate in the trade policy community. New trade statistics are being calculated and estimated. These calculations are based not on the traditional gross figures for exports, including any imported input goods and business services, but on net figures, where only the national value added remains and the imported content has been subtracted.

A recent report from the Swedish National Board of Trade, using IO-tables for the Swedish economy in 1995 and 2005, comes up with a number of new insights. I would like to focus on two of issues that have not received much attention so far in the debate.

First, conventional gross trade figures inflate the value of exports in relation to GDP. The report finds that 33.5% of all Swedish exports consisted of imported goods and services. If one-third of the exports consist of imported inputs that means that only two-thirds of gross exports are actually creating value in Sweden. Only two-thirds of the export earnings contribute to wages and capital returns to Swedish workers and investors. Consequently, when official statistics claim in 2005 that the export share of Swedish GDP was 49%, this is exaggerating the contribution of exports to the Swedish economy. The real contribution is the value added of the exports, which is only two-thirds of 49%, i.e. 31%. It is more accurate to say that approximately one-third of Swedish GDP is generated by foreign demand, not half of it. Thus, Sweden is – like other nations with a similar level of economic development and integration into international supply chains – considerably less dependent on exports than we normally think we are.

Contributing one-third to GDP is still far from peanuts, and we must not look at export figures in isolation, disregarding all the beneficial direct and indirect effects exports have on the domestic economy. The right conclusion is not that exports are “overrated” but rather that imports are “underrated”. Without access to vital imports, export firms cannot thrive. As can be seen in Figure 1, vertical specialisation increased in Sweden between 1995 and 2005 – meaning Swedish exports became “less Swedish” – at the same time as the contribution of net exports to the Swedish economy increased. This positive development would probably not have been possible without Swedish firms taking advantage of the international division of tasks.

Figure 1. Swedish vertical specialisation and value of net exports.

Source: Swedish National Board of Trade

“Made in Sweden” can thus appear an anachronism. It is however not equally anachronistic for all sectors. National aggregates hide huge sectoral differences, ranging from sectors with very little import content to sectors where almost all production consists of imports. Generally, production of services requires fewer inputs than production of goods and, as a consequence, services are also less dependent on imported inputs. This leads me to my second point.

In our study (National Board of Trade 2011), we find that the vertical specialisation of sectors range from 15% (electricity, gas and water) up to 90% (petroleum), whereas for services the range is between 8% (finance and insurance) to 30% (transports and storage). This means that the “importance” of some sectors – the value exports from this sector add to GDP – is overestimated whereas some sectors are underestimated. We should of course be careful to draw too far reaching conclusions regarding individual sectors as all sectors are interconnected and dependent on each other. One thing is clear though and that is that exports of services is contributing more to the economy the gross trade data imply.

Figure 2 illustrates our estimates for this. In conventional gross terms the share of services exports from Sweden in 2005 was 29%. Subtracting the import content of Swedish exports and only looking at the remaining net figures (value added being created in Sweden), we find that the share of services increases to 36%. That is a sizable difference, but that is not all. This covers only services trade via GATS mode 1 (cross border trade). We have an estimate of the value added of Swedish tourism exports, which is essentially identical to mode 2 (consumption abroad) in GATS terminology. These are sales not included in the exports statistics by Swedish retailers, restaurants and hotels to tourists and estimated to constitute almost 4% of total Swedish exports in net terms. Add that on top of the 36% services exports mode 1 and the figure for services exports is now almost at 40%.

Figure 2. Service exports as a proportion of all exports from Sweden in 2005.

Source: Swedish National Board of Trade

Nevertheless, this is still likely to underestimate the importance of services exports. There is service trade delivery via mode 3 (commercial presence), selling services abroad using affiliates.1 The WTO estimated in 2006 that the value of sales via commercial presence was as great as the value of sales via cross-border trade. If we add these services sales to the exports figures, the share of services in total foreign sales goes up.2 The magnitude of this depends on how much of the affiliates' profits are repatriated to Sweden, a figure we do not have.

A highly relevant concept in this context, which we still have only a vague understanding of and where much more research is called for, is servicification. Not only do goods-producing firms use services as inputs in their production, they also sell many services. This naturally applies to exports just as well as sales via affiliates. As we have shown in another study at the board a large part of the sales of goods producing firms consists of logistics, maintenance, insurance and financial services and other services needed to use the good. This is not always properly recorded in export statistics, but often “disguised as goods” in the figures. Even in gross terms services exports is thus underestimated.

Adding affiliate sales of services abroad and the blurring concept of servicification to the 40% services exports share makes it possible, perhaps even likely, that the majority of Swedish firms net foreign sales consists of services, not goods. This probably explains a large part of the gap between the share of services in the domestic economies and the share of services in world trade. There still is a big gap though as services are still much less traded internationally than domestically. The right conclusion should be that services are not only more traded but also more tradable than we think. This knowledge should encourage policymakers and spur them to put more efforts into further liberalising conditions for services trade.

As Pascal Lamy (2011) claims, some trade tensions might be based on a misunderstanding of the trade statistics. If our view of world trade is to some extent distorted, then our policies risk becoming so too. Value-added trade statistics tell another, more accurate story about who is actually exporting what in the world economy and how much they gain from it. If better statistics can contribute to better policies, then the ongoing efforts to calculate value-added trade measures should be encouraged.


Hummels, David; Ishi, Jun and Yi, Kei-Mu (2001), “The nature and growth of vertical specialization in world trade”, Journal of International Economics, 54(1).

Lamy, Pascal (2011), “Made in China tells us little about global trade”, Financial Times, 24 January.

Miroudot, Sébastien and Ragoussis Alexandros (2009), “Vertical Trade, Trade Costs and FDI”, OECD Trade Policy Working Paper 89.

National Board of Trade (2010), “Servicification of Swedish Manufacturing”,Stockholm, Sweden.

National Board of Trade (2011), “Made in Sweden? A new perspective on the relationship between Swedens exports and imports”, Stockholm, Sweden

UNCTAD World Investment Report (several years)

WTO (2007), World Trade Report 2007: Six decades of multilateral trade cooperation: What have we learnt?


1 The last mode of delivery is mode 4, i.e. temporary movement of people – which in the case of Swedes working temporarily abroad to supply services – probably affects this only in a negligible way.

2 But then, to be fair, we should add the sales of goods from the affiliates too and get a complete foreign sales figure. That is probably not decreasing the relative importance of services. UNCTAD estimates that 63% of all ingoing FDI assets in the developed countries were in the services sector, which can lead us to believe that the great majority of foreign sales by affiliates is in services and not goods. In terms of how much value affiliate sales contribute to the Swedish economy we must recognise that a large part of the value added of these activities do not end up in Sweden but rather as wages to foreign workers and payments for locally sourced inputs. Only the (not locally reinvested) capital returns for these investments end up in Sweden. How much that is, we do not know.