David Primack (Executive Director, ILEAP*) @DavidPrimack
11 November 2016
Despite improvements in the collection of services trade data over the past 15 years, in many low-income and least- developed countries (LICs) the macro- and micro- level services data needed for meaningful economic analysis simply do not exist. This acts as a fundamental roadblock to having informed services trade negotiations and to using services trade policy to leverage services for inclusive growth and structural transformation.
The relative paucity of services and services trade data has contributed to obscuring the role that services have increasingly been playing, alongside agriculture and manufacturing, in the process of structural transformation. While emerging research such as ODI’s Supporting Economic Transformation initiative and WIDER’s Industries Without Smokestacks project is helping to advance a more analytically rigorous understanding of the interactions different service sectors have in the transformation process, such efforts continue to be hampered by a number of services-specific data limitations.
Unlike trade in goods, where a single document provides an internationally recognised product code and an indication of the country of origin and destination, as well as a transaction value, collecting service trade statistics is a highly subjective undertaking, prone to inaccuracies and a general dearth of availability. This is particularly the case for LICs, though the challenges prevail in other developing and even developed countries. These challenges are directly related to the nature of services trade and the absence of a physical item (and/or sometimes a payment) crossing a border where national authorities can track, count and record it.
Deficiencies in the measurement and availability of services trade data were flagged by negotiators during the WTO’s Uruguay Round GATS negotiations and continue to hinder services negotiations the world over (e.g. Lipsey, 2006; Magdeleine and Maurer, 2008; WTO, 2010). However significant improvements in services trade data collection have been achieved since 1994. This includes the UN Statistical Commission’s publication of the Manual on Statistics of International Trade in Services (MSITS) (in 2002, revised in 2010), which provides guidelines and recommendations for best practice on how to use and develop sources to measure international trade in services. Four international sources now provide services trade data: the UN Services Database (UNSD); the WTO/UN Conference on Trade and Development (UNCTAD)/International Trade Centre (ITC) Services Database (WTOSD); the OECD Trade in Services by Partner Database (TISP); and the World Bank Trade in Services Database (WBTSD).
While these improvements have enabled better analysis, including in LICs, they have yet to fundamentally address a number of core deficiencies that impede more informed services trade negotiations and policy-making.
The first relates to the source of services trade data and the mismatch with how services trade negotiations are organised. In GATS-based negotiations, services are delineated by the different ways in which they are traded – e.g. online (mode 1 or ‘cross-border’), by consumers visiting the ‘exporting’ country (mode 2 or ‘consumption abroad’), by investment flows establishing an operation that provides services outside the home country (mode 3 or ‘commercial presence’) or by individuals operating outside their home country on a temporary basis (mode 4 or ‘presence of natural persons’). However, services trade data are sourced largely from the balance of payments (BoP), which records transfers of money across borders.
This results in a number of shortcomings. First, in instances where a local consumer pays funds to a locally established foreign affiliate services provider, these payments do not cross a border and are not captured. Investment-related services trade is effectively left out of BoP-based statistics. In that the WTO has estimated such flows to comprise over 50% of global services trade (though surely less so in LICs), services negotiators and policy-makers start with (at best) only half the picture.
Second, when it comes to movement of persons, international services trade statistics utilise proxies in the form of labour-related flows (i.e. compensation of employees, workers remittances and/or migrant transfers). While these provide rough estimates, these data include the activities of permanent migrants and workers outside the services sector, whose work may also not be temporary in nature. This is likely to result in overestimations (offset, however, by the prevalence of the informal sector; see below).
Third, for cross-border services, while available statistics include elements of transportation services, communications, insurance and banking, as well as royalties and licence fees, they generally omit e-commerce transactions (notably where the product is both procured and delivered online). In that e-commerce represents a growing potential delivery channel for LIC service providers (Frost & Sullivan  for example estimate the African e-commerce market will grow by 40% annually over the coming decade), this leaves another key dimension of developing country services trade profiles out of the view of negotiators.
Lastly, with tourism being a major source of LIC services exports, the shortcomings in measuring services consumed abroad can have significant distortionary effects on policy-making and negotiations. For example, trade statistics on tourism generally rely on the travel account under the BoP, which includes not only services but also the purchase of goods by tourists. It also excludes international airfares, which are counted under transport services.
While problematic in their own right, such shortcomings do not even begin to take into account more recent international trade patterns, notably in terms of indirect services trade (i.e. services embodied in goods) in the context of regional and global value chains. Here, trade in value-added statistics based on national input/output tables, are increasingly being deployed, such as those found in the OECD-WTO Trade in Value-Added (TiVA) database, the World Input-Output Database (WIOD), or the Eora multi-region input-output table (MRIO) database. Unfortunately outside of MRIO, few LICs are included in the country coverage, and where available in MRIO, the underlying data tends to be dated and of questionable quality. The SET data portal highlights a number of such services-related indicators of relevance for analysing economic transformation.
Other core deficiencies that preclude the use of existing services trade data to support services negotiations and policy-making relate to the absence of information on bilateral flows (i.e. what partner is the trade happening with?) and sectoral disaggregation (i.e. exactly which services are being traded?).
In recent Department for International Development (DFID)-supported research, Shingal (2015) reaffirms that the coverage of services trade, both aggregate and at the sector level, remains a challenge for least developed countries (LDCs) and LICs, especially vis-à-vis their bilateral trade flows. The latter is particularly so because, in the absence of LICs & LDCs reporting their bilateral flows, partner ‘mirror flows’ are used as a substitute. For example, when UNSD reports Tanzania’s commercial services exports to the UK, these figures are in fact what the UK reports as commercial services imports from Tanzania. While such standard techniques are helpful in filling some gaps, they do not address the gap in South–South services trade flows (as there are no mirror data to use). This has particularly adverse implications for South–South regional services integration: policy-makers and negotiators have virtually no data about the services that flow between the parties to the negotiation. It is perhaps unsurprising then that the private sector is hard-pressed to identify negotiated services outcomes that have a meaningful impact on their business.
Another important phenomenon affecting the availability of services trade data in LICs is the prevalence of the informal economy. To the extent that significant cross-border transactions happen outside the formal sector (e.g. distribution services, personal and professional services), these go unrecorded in the BoP. Similarly, where services operators tend to be micro and small enterprises, a tendency has been observed for successful services firms to export ‘under the radar’, often to avoid paying taxes (e.g. VAT) (Primack and Kanyangoga, 2014).
The low levels of data reliability add additional texture to these challenges. In comparing LIC & LDC services trade data in UNSD, Shingal (2015) identifies high variability in the recorded sectoral coverage across years, as well as at times significant year-on-year variation. These, he suggests, may point to weaknesses in the quality of data collection and transcription/coding. One commonly cited example is the improper handling of exchange rates, where in the context of relatively low overall volumes a few mis-recorded transactions can significantly distort an entire year of data.
LIC services policy-makers and trade negotiators have thus been left to operate in relative darkness. More often than not, the determination of offensive and defensive negotiating interests is left to the realm of anecdote, intuition and, at best, the input of a few (hopefully representative) stakeholders on the barriers they may be facing in external markets they are contesting (or would like to contest). The same goes for identifying binding constraints in the domestic market that may undermine firm competitiveness and limit productive and exporting capacities. While case studies in specific sectors and countries can help, and indeed remain essential even in the best of data scenarios, alone they cannot substitute for reliable, adequately disaggregated services trade statistics for informing services policy-making and negotiations (at both the macro country level and the micro firm level).
Improving the state of services trade data in LICs is a time- and resource-intensive endeavour, but one very much in the realm of the possible. Of note, securing results here need not require operating at the frontier of international best practice. Building on the aforementioned DFID-supported research, Holmes et al. (2016) suggest a sequenced process for improving the collection of services trade data based on the practices of other developing countries that have performed well in this realm. Such practices include, inter alia, undertaking a needs/capabilities assessment, implementing enabling legislative and institutional provisions (including strict confidentiality for reporting firms), use of multiple data sources – in particular targeted firm-level surveys – and securing external technical assistance and capacity-building. Finally, ensuring any such efforts are properly embedded in national planning and budget processes will help promote sustained improvements over the longer term.
 While Foreign Affiliates Trade in Services (FATS) statistics track such investment flows, this is a still-novel and complex methodology, and as such remains largely the purview of advanced economies. BoP data do capture some mode 3 data related to constructions services.
 The use of thresholds under which items need not be declared can exacerbate under-reporting.
 Tourism under the Tourism Satellite Account represents an alternative framework.
Holmes, P., J. Rollo and A. Shingal. (2016). ‘Toolkit: Improving services data collection in LDCs & LICs’, ILEAP, CUTS International Geneva and CARIS: Toronto, Geneva and Brighton.
Primack D. and J.B. Kanyangoga. (2014). Operationalizing the LDC Services Wavier: Rwanda Country Assessment. Mimeo
Shingal, A. (2015). ‘Identifying good practices in services trade data collection in LDCs/LICs’, ILEAP, CUTS International Geneva and CARIS: Toronto, Geneva and Brighton.
* A number of the studies cited emanate from the DFID-funded Trade Advocacy Fund (TAF) project Support to Enhance Development of Trade in Services Negotiations, led by the author. A host of project publications and services-related resources, including on services trade data, are available at www.tradeinservices.net. The author would also like to acknowledge the comments and support from Dirk Willem te Velde and Neil Balchin.
Photo credit: Jonathan Ernst/World Bank