Services trade statistics by modes of supply
Data extracted in March 2019
Planned article update: June 2020
This article describes the Eurostat pilot model methodology used to derive data on international services statistics by modes of supply (MoS). The results in this article were derived by applying the Eurostat pilot model methodology and are not official statistics. This methodology has been applied to international trade in services (ITSS) data from the EU-28 and the EU Member States, for the reference year 2015.
Services data broken down by MoS include both international trade in services (ITSS) statistics and foreign affiliates statistics (FATS). Data users greatly appreciate the statistics on the international supply of services by mode, as it makes it possible to analyse the service supplier’s territorial presence at the moment of a transaction.
Statistics on total trade in services by modes of supply aim to answer the question of how and where services are supplied to foreign customers. These statistics can help better understand the drivers and consequences of international trade in services. Although statistics on services by modes of supply are important for different user groups (policy-makers, trade negotiators, researchers, etc.), there is a scarcity of relevant data, which makes it difficult to properly assess the international supply of services. Eurostat has been working with Member States since 2014 on a model to estimate MoS. In this article, updated results from this work are presented using estimates and experts’ opinions. The data presented in the article are not official statistics and should be considered as being experimental in nature. To produce better-quality and more comparable data, Eurostat continues to work on further developing and better aligning the model to accord with empirical evidence. The current model and applied methodology are described in detail in the section ‘Eurostat pilot model methodology’.
Trade in services by modes of supply
The General Agreement on Trade in Services (GATS), the first multilateral agreement to cover trade in services, defines trade in services as the supply of a service through any of four modes of supply, outlined in the bullet points below.
- Mode 1 - cross-border supply: from the territory of one country into the territory of another country;
- Mode 2 - consumption abroad: in the territory of one country to the service consumer of another country;
- Mode 3 - commercial presence: by a service supplier of one country, through a commercial presence in the territory of another country. The FATS framework is designed to provide information on the activities of enterprises located in foreign markets;
- Mode 4 - presence of natural persons: by a service supplier of one country, through the presence of natural persons of that country in the territory of any other country.
For example, legal services may be supplied to the client (customer) by four separate modes:
- legal advice is provided to the client abroad through electronic channels (mode 1)
- the client from abroad comes to the lawyer’s office (mode 2)
- the lawyer establishes an affiliate abroad to provide legal services (mode 3)
- the lawyer goes abroad to provide the legal service to the client (mode 4).
In practice, services can be produced, distributed, marketed, sold and internationally delivered through a combination of modes. For example, computer services can be delivered either by electronic means (mode 1) and/or by natural persons travelling abroad to provide those services (mode 4). For computer services, we have used proportionality assumptions to allocate the services flows. For example, we assumed that 75 % of computer services were traded by mode 1 and 25 % by mode 4. See the section ‘Applied methodology’ for more details. In further MoS research, general proportional assumptions could be replaced by actual shares resulting from surveys of individual countries.
Exports of services exports are more competitive than exports of goods
EU trade competitiveness can be described by comparing the ratio of net trade (exports minus imports) to total trade (exports plus imports). Table 1 presents these results for goods and services. The EU’s trade competitiveness ratio for total services (taken into account all four modes of supply) was 5.3 %, while for goods it was 3.9 %. This indicates that trade in services may result in a relatively greater surplus than goods trade.
The balance of trade in total services (taking into account all four modes) produced by the EU-28 amounted to a surplus of EUR 232 billion. The 'ITSS balance' (total exports minus imports from ITSS data) had a surplus of EUR 139 billion, and the Services without mode 3 (ITSS data excluding goods and including distribution services) had a surplus of EUR 36 billion (Table 1).
By modes, the largest surplus was recorded for mode 3 (EUR 197 billion), with a proportion to total trade (exports - imports/exports + imports) of 7.3 % (Table 2). For mode 1, a trade surplus of EUR 27 billion was registered, with a proportion to total trade of 2.2 %. The results also show that mode 3 accounts for more than 80 % of the total surplus; modes 1, 2 and 4 taken together generate a surplus of EUR 36 billion while mode 3 alone has a surplus of EUR 197 billion.
For mode 2, a trade deficit was registered as EUR -13 billion (Table 2). Mode 2 mostly comprises travel (including all services provided when the client is travelling, e.g health services, education services and other services). However, it also includes manufacturing and repair services and some transport. Although travel registered a surplus in ITSS, when the other components for mode 2 are taken into account (see Table 3), then a deficit was registered. A relatively significant trade surplus was recorded for mode 4 with a ratio to total trade of 12.5 %.
It should be noted that the above results should be viewed with some caution, as the models are still under development.
EU-28 exports and imports of services in 2015 broken down by modes of supply
Figure 1 presents the services exports from the EU to countries outside the EU broken down by four modes of supply. Mode 3 exports data come from outward FATS statistics (OFATS). In this work, Mode 3 exports were estimated using OFATS data from the economic activities of NACE divisions and sections C33, D to N, and P to S. The partner was all countries outside the EU (extra-EU). Additionally, an estimate for ‘distribution services’ was included in mode 1. For a detailed description of methodology and sources, please see section 'Applied methodology'.
When comparing the EU’s total exports of services (all four modes), to the exports performed through three modes (modes 1, 2 and 4 from ITSS), the amount increases nearly threefold — from EUR 863 billion (sum of modes 1, 2 and 4) to EUR 2 315 billion (also including exports via foreign affiliates). In other words, about 63 % of the EU-28’s total services are supplied via mode 3 (mode 3 has been calculated: EUR 1 451 billion EUR, as a percentage of total exports of EUR 2 315 billion, see Figure 1).
Figure 2 presents EU imports from countries outside the EU, broken down by the four modes of supply. The mode 3 imports were estimated using inward FATS data, from the economic activities of NACE divisions and sections C33, D to N, and S95.
When including the ‘imports’ through foreign affiliates (mode 3), the imported services from countries outside the EU increase from about EUR 827 billion (sum of modes 1, 2 and 4) to EUR 2 082 billion respectively. Hence, mode 3 comprised 60 % of total services imports to the EU-28 (Figure 2).
Mode 1 plays an important role in the supply of services due to the development of new technologies. In Figure 3, the exports of EU-28 services channelled through mode 1 are broken down by service item. The largest contributors are: ‘other business services’ accounting for 27 %, ‘transport services’ accounting for 20 %, ‘telecommunication, computer and information services’ accounting for 14 % and ‘financial services’ also accounting for 14 %.
As already mentioned above, trade in services should not be considered only in terms of cross-border trade; affiliates established abroad should also be taken into consideration. For services supplied through mode 3 (commercial presence), the following service items make the largest contributions: ‘financial services’ account for 19 %, ‘distribution services’ (wholesale and retail sales) account for 19 %, ‘insurance and pension services’ account for 15 %, while ‘other business services’ (including R&D, accounting, legal and engineering services) and ‘telecommunication, computer and information services’ each account for 14 % (Figure 4).
Figure 5 presents data on the volume of exports of services by mode and service item. This shows that ‘other business services’ (including R&D, accounting, and legal services) have the top ranking, with nearly EUR 440 billion exported to countries outside the EU (Figure 5). They are followed by ‘financial services’ (EUR 366 billion) and ‘distribution services’ (EUR 332 billion). In contrast, export volumes of ‘manufacturing services’ and ‘maintenance and repair services’ are marginal. Please note that any services that have been supplied by manufacturers (NACE divisions 10 to 32) through their commercial presence (mode 3) have been excluded from the current study.
For services imports (Figure 6), ‘other business services’ again took the highest ranking, with EUR 493 billion in annual import volumes. They were followed by ‘distribution services’ (EUR 453 billion) and ‘telecommunication, computer and information services’ (EUR 318 billion).
For the share of each mode per service item, there is a lot of variation across service types depending on the specificities of the service (Figures 7 and 8). In exports, mode 3 dominated for ‘construction services’ (accounting for 87 % of this sector), ‘insurance and pension services’ (89 %), ‘personal, cultural, and recreational services’ (84 %), ‘distribution services’ (83 %), and ‘financial services’ (75 %). This is due to the essential need to establish a commercial presence by creating an affiliate in those specific service sectors. For example, an insurance enterprise must also be able to examine the broken car before considering the claims relating to it, so the presence of the insurance company in the host country in the form of an affiliate is essential.
In imports (Figure 8), mode 3 is the dominant mode for ‘construction services’ (accounting for 87 % of this sector), ‘insurance and pension services’ (84 %), ‘telecommunication, computer and information services’ (76 %), ‘personal, cultural, and recreational services’ (also 76 %) and ‘distribution services’ (73 %). Mode 1 is the dominant mode for ‘intellectual property services’, accounting for 90 % of imports in this item.
EU Member States exports and imports of services by modes of supply in 2015
Services exported through mode 3 are more significant for larger economies such as Germany, France and Spain. Services imported through mode 3 are more significant for some of the Member States which joined the EU since 2004.
For each Member State, exports and imports of services by the four modes were calculated using a modelling approach that combined actual exports with statistics from foreign affiliates. Where values were missing, estimates were used instead. For the Member State results, mode 3 was estimated using economic activities of NACE sections G to S excluding O for exports (outward FATS), while sections G to N and division S95 were used for imports (inward FATS). The trade partner was the rest of the world (world excluding the reporting country), and an estimate of the ‘distribution services’ was included in mode 1. For a detailed description of the methodology and sources, please see the corresponding sections below.
There is considerable variation across Member States in terms of total exports and imports of services by modes of supply. This is mainly due to the differences in traded services across the countries, which is also related to their innovative technologies and expertise in trading in services globally.
For exports (Figure 9), mode 1 accounts for a considerable share in Luxembourg (90 %), Malta (73 %), Romania (64 %), Slovakia (64 %), Cyprus (61 %), Latvia (60 %), Belgium (58 %), Czechia (57 %), Hungary (56 %), Bulgaria (52 %), and Poland and Greece (both 51 %). These results may indicate that those countries are more involved in the innovative technologies used to supply services through electronic channels.
Mode 2 is more significant in Croatia (56 %), Bulgaria, Greece, Poland, Lithuania, Czechia, Portugal, Hungary and Romania (ranging between 33 % and 25 % in each). Some of these countries have a significant tourism sector; tourists visiting these countries increase the relative significance of mode 2 exports.
Mode 3 was significant among some of the larger EU Member States, accounting for 72 % of exports for Germany, 69 % for France, 65 % for Italy, and 61 % for both Spain and the Netherlands. Mode 3 is also significant in Sweden (59 %), Denmark (53 %) and the United Kingdom (51 %). Most of these countries have either large populations or large land areas. The results indicate that these countries are more active in capital movement, establishing affiliates abroad to provide the services through mode 3.
Mode 4 is usually the smallest of the four modes in percentage terms. It is somewhat more significant in Malta (where it accounts for 14 % of exports), and in Finland, Romania and Czechia (where it accounts for 10 % of exports of each country).
For imports (Figure 10), mode 1 is dominant in Malta (where it accounts for 63 % of imports), Ireland (54 % of imports), Luxembourg (53 % of imports) and the Netherlands (45 % of imports).
Mode 2 is significant in Greece (where it accounts for 31 % of imports), Denmark (19 % of imports), Lithuania, Cyprus and Germany (12 % of imports for each).
Mode 3 accounted for 74 % of imports in Romania, 72 % of imports in the United Kingdom and Bulgaria, 69 % of imports in Czechia and 67 % of the imports in Latvia, Portugal as well as Slovakia.
Mode 4 imports are more significant in Malta (where it accounts for 9 % of imports), Finland (7 % of imports), the Netherlands and Belgium (6 % of imports).
Services are traded under the General Agreement on Trade in Services (GATS). Services ‘crossing the border’ are supplied without paying any customs duties ‘at the border’. However, the four modes of services supply of the GATS do not account for the fact that a substantial and increasing proportion of services is embedded in a wide range of manufactured products traded around the globe. What would the value of a smartphone be without its software applications (Cernat, L.)? Services supplied through this 5th mode are often technology intensive, such as innovation, design, engineering and product development, and essential for internationally competitive marketing. Hence, services supplied through that mode, embedded in manufactured products in merchandise exports, pay duties, whereas the same services supplied through mode 1 do not. The proportion of these services is substantial and increasing accounting for 34 % of EU manufacturing exports (Cernat, Kutlina-Dimitrova). For example, recent advances in 3D printing, not yet ripe for large-scale deployment for mass production across all industry sectors, may lead to a shift from trade in finished goods to trade in ‘additive ingredients’ and various manufacturing services. Services already play an important role in the manufacturing industries, helping to move the value chain upwards towards business operations with the highest value added. This concept of services trade may be called mode of supply 5, and is not covered in this article. Provided that the necessary data on the services embedded in manufactured products become available, mode 5 (value added concept) in services trade would be an interesting path for future work.
In estimating services trade flows by mode of supply, the ‘simplified approach’ methodology from the Manual of Statistics of International Trade in Services (MSITS 2010) was applied to the publicly available services data (see more details in ‘Data coverage’). In this article, we have used the ‘Eurostat pilot model’, which is based on the simplified approach.
Eurostat pilot-model methodology
The Eurostat pilot model is based on the MSITS approach, but the following improvements have been made. The services, classified by Extended Balance of Payments Services Classification (EBOPS), were assigned to either one mode or to several modes, based on an expert assessment of how specific service items are most likely to be supplied to consumers (Table 3). For example, for ‘computer services’, we assumed that 75 % are supplied through mode 1 and 25 % through mode 4. These assumptions were based on (i) evidence and data from some EU countries, and (ii) expert opinion. For integrating the statistics on foreign affiliates as mode 3, Eurostat has used a conversion table applying equal proportionality assumptions. Therefore, the results of the modelling and estimates can only be considered as a rough approximation of how services are supplied, and comparisons based on these estimates should be viewed with caution.
The existing guidance (MSITS 2010) acknowledges that available data sources have their limitations, and that — in addition to actual data — compilation should include the use of modelling and estimates. One limitation of this modelling approach is that various combined sources could overlap. For example, the turnover of affiliates abroad also includes exports back to the home country or other countries, which is covered by the flows already included in international trade-in-services statistics.
For services channelled through mode 2 (estimates based mainly on travel services), the value of goods purchased by travellers is estimated at about 41 % in imports and 25 % in exports, and has been excluded from the travel item. The calculation was based on the average of four EU Member States where a travel breakdown by goods and services was available. That average ratio was also applied to the EU-28 aggregate.
This study uses publicly available data from the ITSS domain (based on the EBOPS products classification) and FATS (based on the activities of enterprise affiliates OFATS and IFATS). It also uses international trade in goods statistics. For wholesalers and retailers, we have used trade by enterprise characteristics (TEC) and structural business statistics (SBS) datasets (see 'Distribution services in Mode 1 and Mode 3' below).
For the EU trade calculations, we have used EU-28 data with partner all countries outside the EU (extra-EU). Mode 3 was estimated using OFATS and IFATS data. For the EU trade values, we estimated mode 3 as the sum of the following NACE sections and divisions:
- OFATS: C33, D, E, F, G, H, I, J, K, L, M, N, P, Q, R, S.
- IFATS: C33, D, E, F, G, H, I, J, K, L, M, N, S95.
The production value is used for section G, and the turnover value is used for all other cases (see details for section K in the section below). Due to several confidentiality constraints, for the individual Member State results, we have used as partner the ‘rest of the world’ (world excluding the reporting country) and for mode 3 we used data from the following NACE sections:
- OFATS: G, H, I, J, K, L, M, N, P, Q, R, S.
- IFATS: G, H, I, J, K, L, M, N, S95.
FATS as mode 3
Mode 3 is becoming more significant because ‘services suppliers may choose to set up affiliates as a means of selling their services in foreign markets as an alternative or complement to exporting through other modes’ (MSITS paragraph 5.64). The foreign affiliates’ statistics (mode 3) were mapped to EBOPS main categories by using the developed correspondence table from NACE to EBOPS.
For the Member States results, when the source data were missing in the Eurostat public database due to confidentiality or a lack of availability, simple imputations were made (e.g. proportional allocation).
In particular for OFATS, the partner ‘world excluding the reporting country’ is voluntary. Therefore Mode 3 exports was estimated for Bulgaria, Estonia, Luxembourg, Malta and the Netherlands due to full or partial non-availability of OFATS data.
The mode 3 calculation for the EU-28 was based on EU-28 aggregate data whose availability was almost complete (the main exception was the inward FATS data for NACE Section K, for which values for several countries and the EU aggregate had to be estimated).
While production value (or output value by MSITS) is generally considered a superior measure of the service supply (see MSITS 2010, para. 5.65), in most service sectors ‘turnover’ (or ‘sales’ by MSITS) is equivalent to output and, importantly, more easily available. In this article, for most economic activities the ‘turnover’ value was used for estimating services. The exception is ‘distribution services’, corresponding to wholesale and retail sales activities (NACE section G), for which the production value was used. For inward FATS, the production value was used for ‘distribution services’, as it is already available and considered a more accurate measure of services activity for most purposes (MSITS 2010, paragraph 5.65). However, for outward FATS, the production value data are not collected; only turnover data are available. Therefore, we have estimated the production value for OFATS using the inward FATS ratio of production value to turnover. This ratio ranges from 13 % to 42 %, depending on the country. The weighted average ratio was calculated as 25 %, which means that the production value comprises a quarter of the turnover value of the distribution services. This average ratio was also applied to the EU-28 aggregate to calculate the distribution services in mode 3.
For the EU-28 aggregate, FATS manufacturing (C10-32) activities were excluded from mode 3, as these mainly concern manufacturing (of goods), which does not fall within the scope of the GATS. Only ‘maintenance and repair services’ (NACE C33) was used (estimated by C31-C33). Moreover, to avoid over-estimation of mode 3, the trade of accommodation and food activities (NACE I) was estimated at 50 %, as part of that trade may already be covered by the ITSS travel data.
Financial and insurance activities in mode 3
According to the recommendations of the MSITS 2010, measured output (production) will be identical to sales (turnover) for most service activities, except for three activities: wholesale and retail distribution, financial intermediaries, and insurance. Output is generally considered a superior measure of the service supply (see MSITS 2010, para. 5.65). However, as regards financial and insurance activities in FATS (NACE Section K), there are several practical limitations as described below.
In outward FATS (Mode 3 exports), for Section K ‘financial and insurance activities’ only the turnover value is available. Therefore, we used the turnover variable also for Section K. For insurance services, this means generally ‘premiums paid to be insured’; the ‘production value’ (the ‘margins’) would be a better estimate for services supply, however, it is not available.
In inward FATS (imports), the following variables were used for activities in Section K:
- for K64, the production value is available and was used,
- for K65 and K66 the turnover value is available and was therefore used.
Furthermore, the availability of data for Section K in inward FATS is generally limited. Therefore, for several countries, Section K inward FATS values were estimated using proportional assumptions, both for the invididual countries' estimates (partner: 'world excluding the reporting country') as well as the EU28 estimates (partner: non-EU member countries).
As the exports and imports have been calculated in different ways and several estimates used, comparisons between exports and imports should be viewed with caution. Furthermore, because the insurance and financial services for modes 1, 2 and 4 have been estimated using a production value measure, any comparison between modes should also be viewed with caution.
Distribution services in mode 1 and mode 3
Distribution services are defined as the distributive services provided by wholesale and retail trade industries. In the 2008 SNA, wholesalers and retailers are defined as entities that purchase and resell goods with no, or only minimal, processing (for example, cleaning and packaging). Although most distribution services of wholesalers and retailers are excluded from EBOPS, they are included in W/120, which is used in GATS negotiations (MSITS para 3.53).
Distribution services are an important element of the international supply of services both in mode 3 (FATS statistics) and mode 1 (ITSS statistics). In mode 3, distribution services have been estimated by the production value of section G, wholesale and retail trade (please see the section ‘FATS as mode 3’ for details).
However, the part of distribution services that should be allocated to mode 1 is not captured by the ITSS data. Instead, the distribution service value is captured in ITGS, as this value is included in the total value of traded goods crossing the border. Hence, the margins of wholesalers and retailers are generally included indistinguishably in the value of the goods sold.
For the results in this article, distribution services were estimated using merchandise statistics from NACE Section G. We analysed the TEC dataset in Eurostat’s reference database. We estimated the trade margins by using the SBS datasets, while calculating the proportion of the value ‘Gross margin on goods for resale’ in the total turnover value. We applied the trade margin separately for each country: for the EU-28 we used the weighted average, which was calculated as 21 %. Due to limited data availability, we have used the same trade margin percentage for both imports and exports.
Eurostat continues to work with the Member States to improve the model. There are several areas where further improvement is needed. Below we present the most important ones.
Double counting between FATS and ITSS
Double counting may occur in the total trade values. For example, if a foreign affiliate exports to a third country (or to the country of residence of the parent company), the service will be included in the turnover of the affiliate (mode 3) and may also be captured by the ITSS data as exported services (see also MSITS 2010, para. 5.8). Hence, if the turnover of the affiliates’ statistics also includes exports, this will result in double counting, as this value is already captured by the ITSS. In particular, any exports to the affiliate back to the country of origin are not exports.
Value of goods
The value of goods is also included under some service items — e.g. travel, construction, and maintenance and repair items. However, for some of these items, it was not feasible to identify the goods and services separately. In this work, the goods component was excluded only from the travel item, based on information from several EU Member States, which provided breakdown data to Eurostat.
Use of empirical data
For better-quality results, more empirical data will be needed. The long-term goal is to replace the hypothetical distributions by ‘real’ data, when they become available. More refined allocation from NACE to EBOPS should be developed in the future. Furthermore, for mode 3, the services part of the manufacturing activities should also be included in the modes-of-supply model.
Further analysis in our study will follow a time series of the services trade by modes. It could be the case that sales abroad by mode 1 become more significant than sales via foreign affiliates. In addition, future studies could also contain an analysis on job creation via mode 3 in those sectors where mode 3 is significant.
Source data for tables and graphs
Statistics on the international supply of services by mode of supply are being developed primarily to meet the needs of the General Agreement on Trade in Services (GATS) trade negotiators and analysts. Statistics are required to support negotiations and to monitor the impact of services trade agreements.
Statistics on the international trade in services offer a good picture of trade flows between countries, based on the services classification. International trade in services covers services transactions between residents and non-residents. The international trade in services statistics are first collected, and then, after adjustments, used for balance-of-payments purposes. The service transactions are mostly compiled according to the residency of the seller and the purchaser, without any distinction as to how the service is actually supplied. While the services classification offers valuable information on the types of services traded, it is also important to identify where and how the services are supplied by considering the location of the supplier and customer (Table 4). This is an extended dimension of international trade in services, also referred to as international supply of services.
Services differ from goods in respect of the immediacy of the relationship between supplier and consumer. Many services are non-transportable, i.e. they often require the physical proximity of supplier and customer (e.g. accommodation services).