While trade debates focus on tariffs and goods, services have quietly defied deglobalisation and become an increasingly important driver of growth. Services trade not only generates productivity spillovers and labour-market gains, but it is also where future policy tensions could emerge. Yet services trade remains under-measured and under-studied. This column introduces the new Bilateral Trade in Services (BiTS) database, which tracks services flows between countries over time and at the detailed sector level. It documents a declining distance elasticity for services trade in recent years, which is driven by the growing importance of less distance-sensitive service categories.
Services are playing an increasingly prominent role in the global economy. Unlike goods, services trade has continued to grow at roughly the same pace as before the Global Crisis, even as goods trade has plateaued (Baldwin 2022, Baldwin et al. 2024). This resilience partly reflects growth becoming more services-led, not only in advanced economies but also in developing economies (García Guzmán et al. 2024, World Bank 2024). It is also supported by an unbundling of service delivery from physical presence, enabled by digital platforms and cloud computing (OECD‑WTO‑IMF‑UN 2025). In the debate about global trade imbalances, services trade is an often-overlooked component with countries that run persistent goods deficits often running surpluses in services (Boz et al. 2019). However, data for bilateral services flows remain sparse and fragmented. To address this issue, we introduce a new research database on bilateral services trade (Li et al. 2025).
A new database of global services trade flows
The new Bilateral Trade in Services (BiTS) database provides the broadest coverage of officially reported bilateral services trade to date. It tracks bilateral services flows for up to 245 countries and territories across 12 major categories (26 subcategories) from 1985-2023 (Figure 1), all harmonised under classifications consistent with the Sixth Edition of the IMF’s Balance of Payments Manual (BPM6). It consolidates data from multiple sources. Its backbone is the OECD-WTO Balanced Trade in Services database (BaTIS; OECD 2025), complemented by data from the UNCTAD-WTO, UN Comtrade, and Eurostat. To ensure comparability, a concordance converts older BPM5 classifications into BPM6, and a hierarchical reconciliation procedure merges overlapping sources using proportionality assumptions and judgement, as documented in Li et al. (2025).
The result is a research database that brings together officially reported bilateral services trade from a range of widely used sources. For some advanced economies, data extends back to 1985, and for most major economies, BiTS captures virtually the full value of multilateral services exports and imports from the early 2000s onward. By 2010, it accounted for roughly 90% of global services exports (Figure 1), ensuring that missing flows are economically small. Coverage is most complete for advanced economy pairs, with more than half of all possible bilateral flows observed post-2010, while coverage for emerging markets is narrower but improving over time. This makes BiTS suitable for empirical analyses of bilateral services trade patterns that require broad country or time coverage. It is designed to complement, not replace, other datasets such as BaTIS and the International Trade and Production Database for Estimation (ITPD-E; Borchert et al. 2021), which offer features BiTS does not — such as, respectively, a balanced trade matrix and domestic production data.
Figure 1 Bilateral Trade in Services (BiTS) database coverage
Note: The bars depict the number of countries and geographic entities for which at least some bilateral trade in total services (category S) is covered by BiTS in a given year. AEs = Advanced Economies; EMDEs = Emerging and Developing Economies; Other = non-country geographic entities (for simplicity referred to as ‘countries’ in the text). The lines show the percentage of multilateral services exports, taken from the World Economic Outlook Database, that is covered by the sum of total bilateral services exports (category S) in BiTS, both for the median country and for the world as a whole.
Do bilateral services flows obey the same forces as goods flows?
To illustrate the value of BiTS, we examine whether services trade is shaped by the same ‘gravity’ forces as goods. Goods trade is governed by well-established and persistent forces. Figure 2(a) represents the global goods-trade network: node size reflects centrality in the network, line thickness indicates flow intensity, and colours represent clusters identified by the Louvain algorithm, meaning countries in the same colour group trade more intensively with each other than would occur by chance. As shown in Figure 2a, larger economies such as the US, China, and Germany dominate the global goods-trade network, and geography shapes clear regional clusters: North America (blue), East and Southeast Asia (red), and Europe (yellow). This reflects the ‘gravity’ forces first documented by Tinbergen (1962): bilateral trade flows are strongly related to economic size and geographic distance — a finding that remains central to modern trade theory (Head and Mayer 2014, Costinot and Rodríguez-Clare 2014).
The services trade network in Figure 2(b) looks different. The US and Germany remain central, but smaller economies such as the Netherlands and Ireland also feature prominently, while China plays only a modest role. Geography matters less: one dominant cluster spans advanced economies across North America and Europe (blue), while other clusters mix smaller services traders from different regions.
Figure 2 The network of trade in goods and services
Note: Bilateral goods trade data taken from the IMF Direction of Trade Statistics (DOTS). Bilateral services trade data taken from BiTS. Charts drawn using the Louvain community detection algorithm. Size indicates the centrality of an economy in the trade network. Thickness of connecting lines reflects the intensity of bilateral trade. Colour indicates membership of a trade cluster based on modularity, allowing for three distinct groupings.
A ‘naïve’ gravity regression relating bilateral flows to distance and GDP confirms this weaker pull. In 2019, GDP and distance explain about half of the variation in bilateral services flows — impressive for such a parsimonious model, but still below the two-thirds explained for goods. Moreover, while the explanatory power for goods has been stable over time, the fit for services has declined from almost 100% to roughly 50%.
To probe this further, we estimate a more structural gravity regression, controlling for importer and exporter fixed effects as well as a range of observable country-pair-specific trade drivers. The estimated distance elasticity of overall services trade falls from roughly -0.62 in the early 2000s to -0.50 in recent years, as shown by dark blue bars in Figure 3.
There are two possible explanations for this finding. First, technological changes such as digital communications may have made it easier to deliver services to remote locations. Second, the composition of services trade may have shifted in favour of categories that are inherently less distance-sensitive. Traditional services such as transport and travel remain highly sensitive to distance, with elasticities close to those for goods, but distance matters a lot less for modern services — including finance, ICT, and other business services.
To test these hypotheses, we perform two additional estimations. First, we estimate gravity equations at the two-digit service category level and then aggregate the distance elasticities using each category’s trade share. The results (light blue bars in Figure 3) closely match the aggregate estimates of the distance elasticity (dark blue bars). Second, we reweight sectoral elasticities using fixed category shares from 2000-2004 (light blue bars). This counterfactual shows that had the composition of services trade remained unchanged, the decline in the aggregate distance elasticity would have been modest. In other words, the observed weakening of gravity in services trade is driven primarily by the growing importance of less distance-sensitive service categories.
Figure 3 Services trade composition and the distance elasticity
Note: Bars show the distance elasticity of aggregate services trade, averaged by period. The estimates represented by the first set of bars (dark blue) are obtained from estimating a gravity equation for aggregate bilateral services flows. The estimates represented by the second set of bars (light blue bars with solid border) are obtained from estimating gravity equations at the two-digit category level, then aggregating using categories’ shares in aggregate services trade. The estimates represented by the third set of bars (lighter blue with dashed border) are obtained the same way but using constant 2000-04 shares as aggregation weights.
Concluding remarks
Services trade is becoming an increasingly central component of the global economy, and its drivers differ in important ways from those governing goods. The BiTS research database is designed to provide a consistent, wide-coverage source for studying trends in bilateral services trade. Recent applications demonstrate its uses. As we show above, evidence points to a weakening role of geographic distance in services, reflecting both digital technologies and a compositional shift toward less distance-sensitive sectors. In other work, we have documented that services trade has been resilient to an ongoing fracturing of global trade along geopolitical lines (Li and Zymek 2025) and that dominant currencies influence pricing and invoicing patterns in services (Li and Meleshchuk 2024). Together these insights underscore that services can act as a buffer for global trade growth. As digitalisation accelerates and liberalisation opens new avenues for growth, identifying the remaining barriers to cross-border services will be vital — and BiTS offers a new resource for this research agenda.
Source : VOXeu
