A service of the

Download article as PDF

The General Agreement on Trade in Services (GATS) is far broader in policy coverage than conventional trade agreements for goods. At the same time, governments are offered more flexibility to tailor their obligations to sector- or country-specific needs. As a result, commitments vary widely across sectors and modes of supply. Health insurance has proved far more popular, for instance, than healthcare services. Surprisingly, governments have been less selective in other policy contexts, in particular bilateral investment treaties (BITs). Many signatories of such treaties, including individual EU Members, have undertaken potentially challenging obligations across virtually all service sectors. Yet, though frequently invoked, BITs do not meet the same standards, in terms of transparency, open (consensual) rulemaking and legal certainty, as commitments under the GATS.

The impact of the General Agreement on Trade in Services (GATS) on health-related policies has been viewed with scepticism by quite a number of sector experts. However, very little has actually happened. At the end of the Uruguay Round, in 1993/94, a strong majority of WTO Members elected not to assume any trade obligations, in terms of market access or national treatment, in relevant sectors. And there has been little change over time. Typically, medical and hospital services (healthcare) are the only major areas that have remained exempted from the plurilateral request-and-offer process initiated in the wake of the Hong Kong Ministerial Conference of December 2005. For lack of interest, no group of “proponents” has emerged in the Doha Round.1

The situation in health insurance is conspicuously different. Like other financial services, insurance services have been included frequently in schedules of specific commitments. This was due mainly to a “late harvest” after the formal conclusion of the Uruguay Round. Negotiations on all financial services – from non-life insurance to asset management – had been extended twice before finally being completed in late 1997. Interestingly, this was at the peak of the Asian financial crisis. While developed countries’ active economic interest certainly was a driving factor at the time, it was also the risk of further destabilisation, in the event of failure, that might have won over sceptical governments in Asia and elsewhere.

Issues related to the particular role of health-related services, which might warrant attention in a trade context, have not been raised under the GATS to date. Governments seem generally aware of existing legal or definitional uncertainties and, more importantly, of the political sensitivities involved. In the absence of egregious violations of current obligations, nobody may want to launch a legal challenge, whether in health or others service sectors. Yet such hesitations may play a lesser role in other contexts – bilateral investment treaties (BITs) – where perceived trade and investment barriers could be contested as well. And while the possibility of at least discussing and, hopefully, clarifying GATS-related aspects in WTO fora exists, BITs defy multilateral scrutiny and the associated exchanges of intelligence and experience.

The Provision of Health Services: Basic Patterns

The funding mechanisms and contractual arrangements underpinning current health systems in OECD countries are very diverse. If there is one common denominator, it is the fact that, except for the United States, all systems are (nearly) universal in coverage.

Among European countries, two types of arrangements may be distinguished. One relies essentially on tax-financed integrated healthcare where virtually all citizens are treated free of charge at the point of delivery (the “Beveridge model” as practised in the UK, Southern Europe and Nordic countries); the other is based predominantly on statutory sickness funds (the “Bismarckian model”), which reimburse the costs of treatment either to patients (e.g. France) or directly to the healthcare providers (e.g. Germany and the Netherlands). The insurers often operate on a not-for-profit basis and are financed via payroll deductions, employer contributions and direct government transfers.

Most people rely on supplemental private insurance to widen and deepen the levels of coverage. In France, this reportedly applies to about 80 per cent of the population. In Germany, participation in a public sickness fund is obligatory for over 90 per cent of the population; about one-tenth of fund members have complementary private insurance. High-income earners may opt out of the fund but are then normally insured in full by private companies. In other countries, most notably Switzerland and the Netherlands, universal coverage is ensured by competing private suppliers. Providers exposed to particular risks are compensated out of an equalisation fund (Netherlands); certain disadvantaged population groups may be directly subsidised.

The provision of health-related goods and services is frequently organised separately from the insurance system. Medical and paramedical treatment, including physiotherapy and the like, is provided in many instances by private practitioners, including in the form of cooperatives, on their own account. Hospitals may be owned and operated by public entities, private non-commercial organisations or profit-oriented companies, including larger hospital chains. The three types may well coexist within the same jurisdictions; in Germany some 50 per cent of hospital beds are accounted for by public institutions, about 35 per cent by voluntary organisations, and the remainder by private for-profit hospitals.

A common facet of virtually all healthcare systems, including those with strong commercial elements, is intensive government regulation and control. There are normally four reasons for this: to accommodate genuine health-related concerns; to pursue social/distributional objectives; to guard against excessive use and contain costs; and to counterbalance various market imperfections associated with the existence of scale economies and information asymmetries. Such considerations generally apply to the full “supply chain”, from the provision of insurance coverage to sales of pharmaceuticals and (ambulatory) medical or (stationary) hospital treatment. Cost and price controls may be complemented by quantitative limitations, for similar reasons (i.e. to prevent oversupplies and/or excessive treatment), on the number of medical practices or hospital beds that would qualify under public insurance schemes. Such restrictions, applied on a regional basis, may also serve to contain supply imbalances between agglomerations and rural areas.

Potentially Relevant GATS Disciplines

The scepticism surrounding the GATS is not least rooted in its unusually broad coverage of transactions and associated policy obligations. According to Article I, the Agreement applies to government measures affecting trade in services which, in turn, is defined to cover four modes of supply. In addition to the conventional concept of cross-border trade (mode 1), these include the consumption of services abroad (mode 2), the establishment of a commercial presence, normally involving foreign direct investment, in a host market (mode 3) and/or the presence of foreign natural persons as service suppliers (mode 4). However, in comparison to its over 60-year old predecessor in merchandise trade, the General Agreement on Tariffs and Trade (GATT), the Agreement’s broad reach and potential intrusiveness are tempered by much flexibility.

Each WTO Member is required under the GATS (Article XX:1) to submit a schedule of commitments. Yet sector coverage and levels of liberalisation are not specified. Moreover, even in sectors in which commitments are undertaken, Members may attach limitations on market access or national treatment under any of the four modes or even completely exempt individual modes from coverage. In addition, individual segments could be excluded from the scope of a sector as defined in the non-mandatory Classification List generally employed for scheduling purposes (W/120).

At the risk of over-simplification, the granting of full market access, in the absence of limitations, may be equated with a government’s guarantee not to employ any restrictions on: the number of suppliers admitted in a particular service sector; their turnover or assets; the number of operations (branches); the number of natural persons supplying a service; the form of legal incorporation; and the participation of foreign capital. The use of needs tests, under which applications for new licences may be assessed, would also be inconsistent with unlimited market access. In turn, full national treatment implies a commitment not to operate measures that, in law or in fact, would disadvantage foreign services and their suppliers vis-à-vis their like domestic counterparts. Discriminatory measures, which would need to be covered by limitations, could consist of differences in taxation, access to subsidies, constraints on land ownership, etc.

The application of commitments remains subject to an important caveat, however: the requirement of likeness. Thus, the laboratory services or X-ray analyses provided by foreign-owned or foreign-based suppliers would qualify for national treatment only if these are like the services produced, and admitted for sale, by domestically-owned labs within the territory of the Member concerned.

The inscription of limitations and the modification of sector definitions may be used for trade-defensive purposes but could also serve typical public policy objectives. However, there is no automatic link. Depending on a country’s regulatory and institutional regimes, relevant objectives might well be pursued via “non-schedulable” measures, including non-discriminatory regulations (e.g. universal service requirements), taxes or subsidies. Thus, if a government wants to ensure a reasonable regional balance in the supply of medical or hospital services, it may introduce access restrictions, for example in the form of a needs test, on new establishments in population centres. Alternatively, it could operate tax/subsidy schemes that favour a more decentralised structure. While the former measures would need to be covered by limitations on market access, should the sector be scheduled, regional variations in taxation or subsidisation would not be inconsistent per se with GATS disciplines.

The existence of access obligations does not affect a government’s ability to choose whatever regulatory objectives it deems relevant. If there are constraints, these are intended essentially to prevent the commercial value of commitments from being undermined, inter alia, by generally applicable measures that are not administered in a reasonable, objective or impartial way (Article VI:1) or by unnecessarily burdensome regulations, in pursuit of a given objective, that nullify or impair the benefits of a commitment and could not reasonably have been expected at the time of its entry into force (Article VI:5). The precise scope of the latter disciplines is still under negotiation as part of the Doha Round.

The recent financial crisis has shown how quickly long-entrenched institutional and financial arrangements can, or rather must, be modified. The changes were normally in the direction of increased government involvement. In some cases at least, they might prove inconsistent with the commitments a Member had scheduled in the sector concerned. Yet, the GATS provides more room for intervention in financial services than in any other service sector. Pursuant to the Annex on Financial Services, Members are permitted – “notwithstanding any other provisions of the Agreement” – to take measures for prudential reasons, including measures “to ensure the integrity and stability of the financial system” (para 2(a)). Such measures must not be used, however, “as a means of avoiding ... commitments or obligations under the Agreement”.

The Concept of “Governmental Services”

The GATS covers virtually all types of service activities, with two exemptions. They apply to air traffic rights and directly related services (Annex on Air Transport Services) and to services provided “in the exercise of governmental authority” (Article I:3). The latter exemption is defined, rather vaguely, to cover “any service which is supplied neither on a commercial basis, nor in competition with one or more service suppliers”. There is no further guidance; terms and concepts such as “public services” or “services of general interest” are absent from the GATS.

In many countries, services that might be considered “public services”, including voice telephony or road transport, are provided by private commercial operators. Their supplies do not arguably fall under the Agreement’s definition of “governmental services”. However, what about hospital and similar services that are provided on a cost-recovery basis by a government entity? Are these commercial transactions, even in the absence of profit-seeking interests? Possibly yes. Otherwise, the drafters would certainly have opted for a different definitional concept.2

And what about the notion of “competition”? For example, does the mere coexistence of public and private facilities already entail an element of competition? In a study for the Canadian government, VanDuzer convincingly refers to the concept of one-way competition, based on the ruling of a WTO Dispute Panel (Mexico – Telecommunications).3 In the current context, this could imply that private suppliers, keen to poach customers with advertisements etc., may well compete with public-sector facilities which, as long as they remain indifferent, would nevertheless retain their “governmental” status.

Such a one-way concept appears particularly appropriate in view of the broad modal scope of the GATS. Consider the case of a country where all hospital services are provided for free by public entities. Nonetheless, some patients may seek treatment abroad – under mode 2 (consumption abroad) – in order, for example, to elude lengthy waiting lists. A wide definition, which equates coexistence with competition, might severely undermine the carve-out in such cases. This could not have been intended, however. In this context, it is also interesting to note that, unlike other GATS provisions, such as Articles II and XVII (MFN and national treatment), Article I:3 simply refers to the existence of competition without requiring “like” services or “like” suppliers to be involved.

The concept of governmental services has been modified for one sector, financial services. Para 1(b) of the Annex on Financial Services explicitly lists activities that are deemed to constitute governmental services, including “activities forming part of a statutory system of social security” as well as “other activities conducted by a public entity for the account or with the guarantee or using the financial resources of the Government”. Pursuant to para 1(c), should a Member permit such activities to be conducted in competition, they are deemed to fall under the Agreement.

The twin criteria of GATS Article I:3 thus have been reduced to its second element, absence of competition. The (non-)commercial nature of an activity no longer matters. In turn, this tends to widen the scope of the carve-out in the services covered by the Annex, which may include health insurance if the above criteria are met (“forming part of a statutory system of social security” or “conducted by a public entity...”). However, as noted by VanDuzer4, an additional issue must be considered in this case: the one-way concept of competition, as proposed before, no longer applies. Even if a public health insurer remains completely indifferent vis-à-vis private suppliers, its services would not qualify as “governmental services” once the latter are allowed by the authorities to compete (para 1(c)). Thus, while the absence of competition already suffices per se to trigger the carve-out, the relevant sector must remain closed to alternative suppliers, regardless of their actual conduct.

Uncertainties remain. In particular, it might prove difficult to determine whether health insurance is part of a “statutory system of social security”. Though there might be no problems in the case of Germany’s Allgemeine Ortskrankenkassen and France’s Sécurité Sociale as far as basic (mandatory) coverage is concerned, the situation may be more complex in other countries. ILO Convention 102 of 1952 lists medical care among the nine branches of social security, but can this provide guidance in countries that have not ratified the Convention (e.g. the United States, Finland and the three Baltic states)? Also, the scope of the second potentially relevant option – “other activities conducted by a public entity” – may prove too limited in some cases. The relevant definition explicitly excludes entities “principally engaged in supplying financial services on commercial terms” (section 5(c)). However, does this really matter?

The carve-out for governmental services is essentially intended to ensure that governments remain free to select, organise and regulate relevant entities as they see fit, regardless of any obligations under the Agreement. The potentially most powerful among these obligations, which is horizontally applicable regardless of the existence of commitments, is the most favoured nation (MFN) principle, i.e. the requirement not to discriminate between like services and service suppliers of different foreign origin. Yet it is difficult to think of any persuasive reasons that would call for public-service functions, whether health insurance or hospital services, to be conferred in a discriminatory manner on suppliers from one particular WTO Member at the expense of others.5 In turn, this also suggests that the carve-out’s modification for financial services, concerning in particular statutory systems of social security, is of limited importance in practice.

There is one scenario, however, in which definitional uncertainties may matter. Imagine a country undertakes commitments in a particular sector, say hospital services, in order to attract international investors and promote the associated transfers of expertise. If the established public hospitals provide these services essentially for free, they can be safely ignored in this context. Otherwise, if they start charging significant fees and poaching patients (e.g. through advertisements), the situation appears less clear-cut. Once these hospitals cross the definitional Rubicon and can be deemed to act on a commercial basis and to compete, any government measure in their favour would define the benchmark for the treatment of all hospitals, including private facilities. The Agreement’s market access and national treatment obligations would henceforth apply across all sector segments. Again, however, this must not be a cause for concern. The GATS is flexible enough to enable Members to avoid such effects, including through tailor-made sector definitions. If there is a challenge, it lies in ensuring adequate consultations among the ministries and agencies concerned and helping them to express sector-specific policies in “GATS language”.

Current Patterns of Health-related Commitments

As noted before, access obligations under the GATS are determined essentially by the specific commitments inscribed in service schedules. In non-committed sectors, Members remain free to operate whatever regimes they deem appropriate, whether complete access bans, unfettered liberalisation or anything between. As indicated before, if there is a constraint, hypothetically at least, it is the MFN obligation.

In terms of commitments made, insurance and hospital services mark the opposite ends of a spectrum. Apart from tourism, no sector has mustered more commitments than (life and non-life) insurance, currently scheduled by over 100 Members. Hospital and other health-related sectors are trailing, with no more than some 60 commitments, representing about one-third of the Membership (Table 1). The result for insurance, in the wider context of financial services, may be attributed mainly to three factors: (i) the high stakes involved in the relevant negotiations which, out of dissatisfaction over the initial Uruguay Round schedules, were extended until the end of 1997; (ii) countries’ self-interest in the availability of core “producer services”, including finance, at competitive conditions; and (iii) the fact that distributional or social policy concerns played no particular role in these negotiations, possibly due to the almost exclusive involvement of ministries and agencies concerned with financial regulation and supervision.

Table 1
GATS Commitments on Health-related Services, October 2009
Number of WTO Members1
Medical and dental services Nurses, midwives, etc. Hospital services Healthinsurance2
Total number of commitments 66 35 58 107
Market Access
Mode 1 Full3 24 (-2) 9 (-1) 22 14
Partial 12 6 1 19
Unbound 30 20 35 74
Mode 2 Full3 49 (-3) 24 (-1) 50 (-1) 31 (-2)
Partial 14 10 5 16
Unbound 3 1 3 60
Mode 3 Full3 19 (-8) 7 (-3) 25 (-9) 23 (-16)
Partial 40 26 24 82
Unbound 7 2 3 2
Mode 4 Full3 0 0 0 2 (-1)
Partial 60 33 54 95
Unbound 6 2 4 10
National Treatment
Mode 1 Full3 27 (-6) 10 (-3) 25 (-4) 41(-5)
Partial 10 6 1 24
Unbound 29 19 32 42
Mode 2 Full3 48 (-9) 24 (-4) 50 (-6) 54 (-10)
Partial 13 10 5 20
Unbound 5 1 3 33
Mode 3 Full3 32 (-13) 21 (-8) 37 (-18) 60 (-37)
Partial 28 12 17 39
Unbound 6 2 4 8
Mode 4 Full3 1 (-1) 2 (-1) 2 (-2) 10 (-1)
Partial 60 31 51 86
Unbound 5 2 5 11

1 EC Members are counted individually.

2 The numbers are approximate since it was not possible to verify in detail whether the classifications used covered health insurance in all cases.

3 Figures in parentheses: reductions in the number of full commitments taking into account horizontal limitations applying across all scheduled sectors.

In contrast, health and other social services did not attract particular interest in the Uruguay Round. Those Members that scheduled commitments essentially did it on their own initiative, apparently without prodding from trading partners or potentially interested businesses. The role of the GATS, if any, has remained confined to that of a “supporting actor”: not a liberalising force, but an instrument to add predictability to existing regimes and, thus, lower the risks perceived by potential market entrants.

On average across all sectors, including those discussed in this article, mode 2 (consumption abroad) is the mode with the highest share of full commitments, about one-half of all entries, which are not subject to limitations.6 This open approach may reflect the view that, since supplier and consumer interact in a foreign jurisdiction, they are largely beyond the scheduling Member’s control. (Nevertheless, interventions are possible via foreign exchange restrictions, visa charges, or the non-extension of otherwise available consumer subsidies.)

In contrast, mode 1 (cross-border trade) has been left unbound frequently. Two considerations may have played a role. In a variety of sectors, governments might have felt that, since supplier and consumer must be simultaneously present to perform a service, including medical interventions, commitments on cross-border trade would be economically meaningless. In other cases, Members may have hesitated to commit on the treatment of services that are produced in other jurisdictions, beyond national regulatory control, in order to then be supplied cross-border. Such hesitations may be particularly strong in regulation-intensive sectors, including many professional services.

Similar considerations may explain why, in contrast to mode 1, virtually all entries under mode 3 (commercial presence) imply access bindings. In many cases, however, these are subject to limitations. Nevertheless, mode 3 is the most commercially important form of transaction, estimated to represent more than one-half of all services traded under the GATS, while modes 1 and 2 account for some 25-30 and 10-15 per cent, respectively.7

Mode 4 (presence of natural persons) has remained economically insignificant by comparison. This may be attributed not only to geographic, cultural and similar barriers, but also to stringent access restrictions. Moreover, the definitional scope of mode 4 under Article I:2(d) of the GATS is essentially limited to self-employed foreign service professionals and to foreign employees of foreign-owned or foreign-controlled companies who provide services in a host country. Foreigners employed by domestically-owned facilities are not covered.

Actual trading conditions may be far more liberal than what Members have “on the books”. In scheduling commitments, governments might have deliberately maintained a margin for future (restrictive) action. This margin is likely to have widened since the conclusion of the Uruguay Round in 1993/94, given that many sectors have been made more accessible. Nevertheless, there are also indications that commitments are not fully complied with. A study examining services policies in so-called transition economies, which account for many recent WTO accessions, found an inverse relationship between the level of a country’s GATS commitments and the openness of its applied regime. According to the authors, one possible explanation is a lack of commercial interest on the part of other Members in the markets concerned and, thus, of incentives to enforce compliance through WTO dispute settlement.8

Public Policy Objectives Reflected in GATS Commitments

The scope of any GATS commitment can be circumscribed in several ways, as indicated before. In order to accommodate sector-specific policy considerations or institutional constraints, a Member may either depart from the widely used Classification List (W/120) or schedule limitations under market access or national treatment for any of the modes.

The United States has used all three options to tailor its commitments on hospital services. First, government owned and operated facilities have been excluded from the sector’s scope, possibly with a view to preventing policies in these segments from serving as benchmarks for the treatment of commercial operators. Second, the schedule reserves the right to prohibit new establishments under mode 3 (commercial presence) should there be no need. Such “needs tests” may be intended, inter alia, to prevent excessive capacity increases and any associated consequences for the length, intensity and, by implication, costs of hospital care. Finally, a national treatment limitation under mode 2 indicates that federal and state governments will not reimburse healthcare expenses incurred abroad. In turn, such restrictions may be deemed to avoid “health tourism” and, by implication, protect the integrity of domestic insurance plans. (Of course, there are more cynical explanations as well.)

The schedule submitted by European Communities (EC 12) at the end of the Uruguay Round neither modifies the sector coverage of hospital services nor contains national treatment limitations for consumption abroad (mode 2). This is interesting against the backdrop of EC-internal rulings concerning the portability of insurance coverage under public health plans. In several cases since the late 1990s, the European Court of Justice has confirmed the principle of patient mobility between the Member States for both ambulatory treatment and, subject to caveats, hospital care.9

The EC schedule does contain limitations under mode 3 (commercial presence), however. In a horizontal section, under market access, the Communities reserved the right, for all scheduled sectors, to subject “services considered as public utilities… to public monopolies or to exclusive rights granted to private operators”. (A non-exhaustive list of relevant sectors and circumstances is added.)10 In a similar vein, concerning national treatment, the schedule stipulates, inter alia, that “[t]he supply of a service, or its subsidisation, within the public sector is not in breach of this commitment”. Further, the sector-specific entry for hospital services is subject to a variety of limitations, mostly referring to needs tests, governing the approval of capacity increases in France, Italy, Luxembourg and the Netherlands. Similar limitations have been inscribed for medical, dental and midwives services in some EC Members.

The commitments on financial services that the USA and EC scheduled at the end of the extended negotiations in 1997 do not provide for any market access for non-life insurance, including health insurance, under either modes 1 or 2. As regards mode 3, the United States inscribed a variety of potentially relevant limitations, mostly specific to individual US states, which may hamper, but not preclude, foreign entry (e.g. citizenship requirements concerning members of the board of directors or restrictions on direct branching into the USA). Similarly, the EC scheduled a range of mode 3-related limitations for individual Member States; they should be read in conjunction with the Communities’ horizontal exclusion concerning public utilities.

A juxtaposition of the EC and US schedules reveals some interesting interpretational differences. They concern in particular the application of the national treatment obligation to services supplied from or consumed in another Member’s territory, i.e. the scope of modes 1 and 2. A case in point is the treatment of subsidies. While the EC has not inscribed any cross-sectoral national treatment limitations for subsidies under modes 1 and 2, the United States and some other WTO Members, including Norway, Liechtenstein and Switzerland, have done so. The EC also omitted from its consolidated EC-25 schedule the subsidy-related limitations that some of its recently acceded Members, including Estonia, had initially inscribed under modes 1 and 2 in their national GATS schedules. The Communities had taken the view that no such limitations were necessary, referring to the “Scheduling Guidelines”, a background document endorsed by WTO Members that is expected to govern the scheduling of commitments.11

Like Article XVII, the Guidelines distinguish between the treatment of services and that of service suppliers. They clearly stipulate that bindings under Article XVII do not require a Member to offer domestic subsidies to suppliers established in another jurisdiction (paras 15 and 16); however, there is no similar clarification concerning the services that may be imported from or consumed abroad. By implication, national treatment disciplines thus do apply to measures affecting the purchase and use of foreign-produced services. (For example, such measures could consist of the non-extension of otherwise available tax breaks to distance-learning courses, ship repairs, medical treatments, laboratory tests and all types of insurance that are supplied cross-border or provided abroad.) It might have been such considerations that prompted the USA as well as Poland, Latvia and Slovenia, prior to their EC accession, to inscribe national treatment limitations for hospital services concerning the non-reimbursement of expenses incurred abroad.

There are indications that the EC’s approach is not cast in stone. A closer look at the recently concluded Economic Partnership Agreement with the CARIFORUM States reveals interesting modifications compared to the Communities’ GATS commitments. First, the EC has narrowed the coverage of health and social services, including hospital services, to “privately-funded services”; second, the commitments under modes 1 and 2 explicitly exclude subsidies from coverage. While the first modification is contained only in the Communities’ EPA schedule, the latter limitation applies horizontally across all sectors and modes committed by the signatories (Article 60(3)).12 However, what are the legal effects of these modifications as long as the EC’s GATS schedule remains unchanged? Concerning the treatment of other WTO Members, the Vienna Convention on the Law of Treaties leaves little doubt. According to Article 34, a treaty “does not create obligations or rights for a third State without its consent”. It might even be argued that the CARIFORUM States have also retained their rights under the GATS (see below).

When Commitments May Need to Be Changed (or Not)

Health services are among those sectors for which it is particularly difficult to anticipate the duration of existing legal and institutional frameworks. Modifications abound – for budgetary, quality and/or social policy reasons. While it might be possible in many cases to accommodate such modifications within the scope of current GATS commitments, there are also instances where commitments may need to be adjusted.

In a number of countries, capacity increases in the health sector (number of physicians, dentists, etc.) are linked, inter alia, to the financial status of the public insurance scheme or to shortages/oversupplies in particular segments or regions. In other words, the number of licences issued might vary significantly from year to year and, possibly, across regions. Such variations can easily be accommodated in a GATS schedule. A case in point is the needs test concerning the establishment of new hospitals inscribed by the USA under market access, mode 3. Similar tests have been scheduled by seven EC Members (Belgium, France, Italy, Luxembourg, Netherlands, Spain and Portugal). Germany and the UK have made the same reservation for medical, dental and midwives services, while Denmark and France reserved the right to regulate the number of doctors, dentists and midwives on a 12- or 18-month basis.

There are discussions in various countries not only about modifications of established regimes, but also about institutional changes in pursuit of various policy objectives (financial, social, etc.). Such changes could entail the extension of insurance monopolies to hitherto excluded population groups in order to improve their situation or, conversely, to tap high-income earners as an additional source of revenue. Thus, the issue has been raised whether Germany’s public sickness fund should be extended to and provide basic coverage for all population segments, including the 10 per cent that are currently not required to contribute.

The extension of a governmental service to additional activities, and the ensuing legal obligations, are not explicitly addressed in the GATS. A closely related scenario, however, is captured by Article VIII:4. A government that envisages granting monopoly rights in areas subject to specific commitments is required to disclose its intention and, at the request of potentially affected Members, negotiate compensatory adjustments elsewhere in its schedule. The relevant procedures are set out in Article XXI of the GATS. It is only after a solution has been found, involving arbitration if need be, that the monopoly may be introduced. Against this background, it appears doubtful whether Members are free to extend, at their discretion, a governmental service segment within a larger sector on which commitments have been undertaken. This might defy the very purpose of such commitments, i.e. to provide a transparent and predictable framework for international trade and investment. As noted by VanDuzer in a similar context, a treaty interpreter must be expected “to divine what the parties intended at the time the treaty was concluded”.13 Nevertheless, not everybody may be ready to concur. Sceptics might wonder whether the existence of explicit rules for one scenario – creation of monopolies in committed areas – and the non-treatment of a parallel scenario – extension of a governmental service – are intended to convey a message: there is no requirement to compensate in the latter case.

The GATS also imposes constraints on the conduct of public monopolies, e.g. in basic health insurance, which seek to diversify into areas where they are in competition, such as supplemental insurance. Article VIII:2 requires the Member concerned to ensure that the supplier concerned does not abuse its position to act inconsistently with existing commitments. Otherwise, these would need to be modified pursuant to Article XXI.

Renegotiations of schedules are very rare, though. The only case completed so far relates to the Communities’ enlargement to EC 25 and does not involve health-related sectors. The dearth of relevant initiatives might be attributed to:

  • the low number of existing commitments, especially in potentially sensitive sectors such as health, and thus, the limited potential for conflict;
  • ongoing liberalisation moves during the past 1½ decades, which have further eroded the substance of “old” bindings;
  • governments’ expectation that trading partners tolerate (mild) infringements, especially in politically and socially delicate cases;
  • country-internal information and coordination problems between the government and the private sector (e.g. the companies concerned might be unaware of the GATS); and, not least,
  • the absence of incentives to proactively renegotiate WTO commitments prior to the introduction of potentially inconsistent policies.

The worst conceivable outcome of a dispute is a recommendation for the Member concerned to bring the measure(s) into conformity with GATS obligations. In the case of continued non-compliance, an arbitrator may have to decide on compensatory new or improved commitments in other areas that are of commercial interest to affected trading partners. The whole process might take one or two years. As distinct from investment treaties, there is virtually no basis in the WTO for aggrieved Members, let alone individual companies, to claim compensation for losses experienced due to the breach of WTO obligations.

Developments in the Doha Round

Given the diversity of (legitimate) trade protection instruments in services, combined with a host of regulatory requirements, it is difficult to develop meaningful indicators of actual or scheduled access conditions. An ambitious study in this respect has been completed recently, based on surveys in 24 OECD countries and 32 developing and transition economies.14 Key trade barriers were rated on a five-point scale according to their levels of restrictiveness in order to then be compared with the countries’ GATS commitments and Doha Round offers. The focus was on the most economically relevant services, in terms of cross-sectoral effects on trade and competitiveness: financial services (including life insurance), telecommunications, retail distribution, maritime transport and selected professional services (accounting, auditing and legal services). The main findings: while the offers would improve the security of access to some degree, they do not involve any element of liberalisation. On average, the best offers are still 1.9 times more restrictive than what currently exists in practice.

A more detailed study of Doha Round developments in financial services confirms this sobering picture. Looking into the offers submitted by 65 Members, counting the EC Members individually, it finds their substance to be “very poor”.15 Only six Members have added subsectors, and in most cases no new business or access opportunities would be granted. In particular, many developing countries are still very far from binding current levels of openness. Countries with potentially attractive markets – Argentina, China, Colombia, Malaysia, the Philippines and Thailand – have not yet submitted any offers in financial services. Nonetheless, this bleak assessment needs to be set against the comparatively meaningful commitments, in commercial terms, achieved in the extended negotiations in 1997. Moreover, as in other sectors, there is still the possibility of improvements being made during the final stages of the Round.

Concerning medical and hospital services, the Round has been even more of a non-event to date. The negotiating momentum was weaker than in any other large sector; it did not even suffice for the formation of a group of like-minded Members to pursue common interests. (In total, some 20 such “plurilateral groups” were constituted in the wake of the Hong Kong Ministerial Meeting of December 2005.) Indeed, in these sectors several participants, including Canada and the EC, expressly confirmed their unwillingness to undertake any commitments (Canada) or to improve on their current schedules. Overall, of the 95 Members that are covered by services offers (as of the end of 2009), only 11 envisage new or upgraded commitments in healthcare services. These are all developing countries.

The implementation of current offers would thus widen the existing gaps between GATS commitments on intermediate (producer) services, including finance, telecom and various business services, and those on consumer-oriented sectors, such as health, education and audiovisual services.

Commitments in Other Policy Contents

There is a strange contrast in the public perception of international treaty obligations. The focus is on developments in the WTO, while relatively little attention is being paid to a rapidly rising number of preferential trade agreements (PTAs), which have recently begun extending to services as well, and an even more dramatic proliferation of bilateral investment treaties (BITs). If the number of disputes were to be used as an indicator of an agreement’s economic significance, the GATS would trail not only the GATT, but also the over 2,500 BITs that are currently in force. No more than a handful of services-related disputes were brought under the GATS between 1995 and 2010, as compared to some 120 cases under investment treaties.16

Preferential Trade Agreements: The EC-CARIFORUM EPA

About 70 preferential agreements have been concluded and announced so far under GATS Article V (Economic Integration). A broad-based study looking into the commitments made on cross-border trade and commercial presence in 40 PTAs found these to be far more ambitious than the respective countries’ GATS schedules and current Doha Round offers.17 This applies across virtually all sectors. Nevertheless, comparable to GATS commitments, producer services have generally drawn more attention than consumer-oriented services. Typically, in terms of commitments made (number of inclusions and levels of openness), healthcare is the worst PTA-performer among all large sectors.

In at least one case, health-related commitments under a PTA are even more narrowly defined than those inscribed under the GATS. As indicated before, the EC’s current GATS schedule, mostly dating from 1993/94, includes commitments on hospital services as captured by the relevant classification number, CPC 9311. For the then 12 Member States, these commitments have no limitations on national treatment under modes 2 and 3. (As indicated before, full commitments on mode 2, consumption abroad, may be understood to guarantee insurance portability under public health schemes if nationals consume like services abroad.) In contrast, the EC’s commitments under the CARIFORUM Economic Partnership Agreement (EPA) reduce the scope of health and social services to “privately funded services”. They thus exclude the public-sector segment which – given comparatively generous insurance regimes on the EC side – might prove commercially attractive for a number of CARIFORUM countries: offering sea, sun and surgery to publicly-funded foreign patients. Moreover, the EPA signatories agreed to exempt subsidies from national treatment across all sectors and modes.

It is difficult to see, however, how preferential agreements could be used to downgrade the parties’ obligations and commitments under the GATS. Provided they meet the relevant conditions in Article V:1 (substantial sectoral coverage and absence/elimination of substantially all discrimination), PTAs permit participants to disregard their MFN obligation in exchanging preferences. Yet, the Article does not provide a basis for extending “MFN-minus treatment” among WTO Members. In any event, the Communities’ approach confirms how cautious governments are in dealing with health services in trade agreements.

Bilateral Investment Treaties (BITs)

Trade and investment issues have traditionally been viewed from different angles and treated separately. However, the GATS has blurred the borderline. Article I:2(c) defines trade under mode 3 to consist of the supply of a service “by a service supplier of one Member, through commercial presence in the territory of any other Member”. This implies that investment-related measures which impinge on such supplies are within the scope of the GATS. The implications are potentially significant, given that mode 3 represents the lion’s share of services trade under the Agreement.

Virtually all WTO Members have concluded and ratified investment treaties, with variations in scope and content.18 They are typically organised in the form of (overlapping) hub-and-spoke systems around major source countries, such as Germany, Switzerland, China and the United Kingdom, which use their own templates. Each of these countries accounts for over 100 treaties. Most of them have no limitations in sector coverage, thus extending across the whole economy, including all service sectors, of the signatory States.

BITs typically contain a range of obligations – including “fair and equitable treatment” and national treatment post-establishment, in some cases even pre-establishment – that do have counterparts in the GATS. Cases in point are the requirement, in scheduled sectors, to ensure the reasonable, objective and impartial administration of measures of general application (Article VI:1) and the commitment to national treatment pursuant to Article XVII. Moreover, there are obligations in BITs, such as a compensation requirement for expropriations, which may affect trade in services but are without direct equivalents in the GATS.

Expropriations conducted in a non-discriminatory manner, according to the principle of due process and serving a public purpose (e.g. terminating commercial activities in environmentally sensitive regions) might be considered GATS-compatible, even if the foreign investor has received no adequate compensation. The property rights usually protected under BITs extend to the enjoyment of intangible assets, including the right to make profits and distribute dividends.19 While GATS provisions such as Articles VI (Domestic Regulation), XVI (Market Access) and XVII (National Treatment) may play a role in protecting an investor against unlawful expropriations in certain circumstances, related BIT disciplines are more specific and, hence, more immediately relevant. Moreover, these disciplines are not enforceable between States only; most treaties also allow for investor-to-State arbitration. The latter element, combined with various GATS-plus provisions, has led observers to conclude that the GATS is less restrictive than standard BITs.20

The 40-odd treaties concluded by the United States cover the pre-establishment phase as well. Consequently, the signatories are committed to extending any liberalisation measures immediately to foreign investors, thus removing leeway for policy experiments. (Recently, other countries have included, to varying degrees, similar obligations in their BITs, e.g. Canada, Finland and Japan.) The sector coverage of the US-promoted treaties is subject, however, to certain reservations. As far as services are concerned, these concentrate chiefly on the insurance, banking and certain transport and communication sectors. Hospital and medical services are not among them. In turn, most of the US treaty partners – composed of LDCs, developing countries and several transition economies – have also scheduled reservations, with differences in focus. In 15 cases, these include insurance services; health services have been exempted only once. The precise scope of these reservations is difficult to ascertain, however, since the focus of BITs is on the assets they seek to protect. Sector definitions are of secondary importance only. Thus, unlike most GATS schedules, investment treaties do not refer to international classification schemes.

Given their particular nature, investment treaties are difficult to reconcile with the criteria for preferential trade agreements under the GATS. Pursuant to Article V:1, footnote 1, such agreements “should not provide for the a priori exclusion of any mode of supply”. However, the scope of investment treaties is confined almost exclusively to mode 3. In areas of BIT/GATS overlap, the MFN clause of Article II of the GATS thus provides for the “multilateralisation” of BIT obligations, whether these relate to national treatment or other disciplines beyond the parties’ obligations under the GATS. There are essentially two possibilities to prevent such effects: The integration of BIT provisions into full-fledged preferential trade agreements or, in the case of self-standing BITs, the listing of an MFN exemption. Interestingly, the latter possibility, which existed only at the date of the WTO’s entry into force or, if later, WTO accession, was used by no more than one-tenth of the current 150-odd Members. Nevertheless, despite its potentially significant role, the “supporting actor” (GATS) has remained off the stage to date: BIT-related complaints have never been raised in the WTO.

The absence of such complaints might be attributed mainly to two factors: First, in quite a number of cases, governments may extend relevant benefits, in particular national treatment with regard to investment grants, production subsidies, etc., to investors from all countries. Thus, there is no scope for friction. Second, should compensation be denied, e.g. in expropriation cases, investors from a non-BIT signatory have little to gain from a WTO dispute. Even if endorsed by a panel, relevant complaints might prove ineffective. As indicated before, WTO dispute rulings are essentially prospective in nature, e.g. calling on Members to bring their policies into conformity with relevant provisions, but do not provide for retrospective compensation.

In any event, affected investors are keen to defend directly what they consider their legitimate commercial interests without involving “their” governments and possibly subordinating their cause to wider policy considerations. Thus, not surprisingly, since the WTO’s entry into force in 1995, which coincided with a sudden surge of new BITs, the caseload under such treaties easily dwarfs the few disputes launched under the GATS.

Compensation requirements under BITs might prove relevant whenever a State redraws the borderline between private and public sector provision of a particular service in favour of the latter. A recent study refers to legislation in the Czech Republic in early 2006 which curtailed the reimbursements of private hospitals under health insurance schemes. Although accessibility to hospitals has improved overall, observers warned that there could be arbitration claims under investment treaties should any of these hospitals be owned by foreign investors.21 More recently, under a BIT between the Netherlands and the Slovak Republic, a Dutch company reportedly challenged its treatment under new legislation which reversed some market-oriented reforms. Under the disputed law, health insurers are required, inter alia, to plough profits back into the healthcare system rather than paying dividends.22 As indicated before, it might prove difficult in such cases to find a violation of GATS provisions.

Observers have argued that “one major disadvantage” of GATS commitments is the difficulty of policy reversals if the commitments are (too) closely geared to the prevailing regimes.23 However, such assessments are somewhat incomplete as long as BIT-related constraints are not taken into account as well. And these constraints have been accepted by a far higher number of WTO Members, intentionally or otherwise, than health-related commitments under the GATS. Though there are flexibilities in BITs, under what might be termed public-interest or public-welfare clauses, their significance would need to be vetted case by case. Moreover, such clauses are unlikely to shield from certain types of claims, including those relating to “fair and equitable treatment”.24

In concluding this section, comments made at a conference on WTO law in 2008 seem particularly apt: “Contrary to the multilateral trading system, no multilateral institution permanently administers BITs or hosts a dispute settlement system to resolve disputes about the interpretation and application of those treaties. The proliferation of BITs has resulted in legal chaos...”25

Concluding Observations

It is difficult to identify common patterns in the GATS commitments on healthcare and health insurance services. While medical and hospital services have been widely ignored, health insurance is among the most frequently committed sectors. Current Doha Round offers would not close this gap, quite the contrary. If there is an element of commonality among virtually all services, however, it is the absence of WTO-negotiated liberalisation, barring some sector- and country-specific exceptions (telecommunications and recent WTO accessions).

Nonetheless, there might be instances where commitments go beyond existing and/or envisaged policy regimes. Potential cases include discriminatory subsidies, in particular with regard to modes 1 and 2, and the non-exclusion, where appropriate, of public sector segments from coverage. Also, there might have been policy changes, possibly including the extension of public insurance schemes into commercially organised segments, in contravention of existing commitments. However, in such socially and politically sensitive cases, affected Members might not want to launch legal challenges under the GATS. In the event of successful complaints, nevertheless, the defendants will certainly prefer compensatory new commitments in less sensitive sectors to the restoration of the “old” regimes.

Governments’ reticence to schedule at least current levels of access under the GATS may come at a cost: international investors may be more reluctant than otherwise to transfer resources and the associated skills and expertise. Yet many countries have insured themselves against such effects as far as mode 3 (commercial presence) is concerned. Under most investment treaties, established foreign suppliers are entitled to full national treatment combined with some additional guarantees for which no equivalents exist under the GATS, such as compensation for expropriation. Yet, the “BIT insurance” is not for free. Some 120 arbitration cases in services since 1995 testify to investors’ resolve to defend their interests under such treaties – and the main targets are developing countries, which account for 90 per cent of the cases.

The proliferation of BITs, coinciding with a stalemate in other international fora, points to a serious lack of policy coherence. Investment treaties are typically prepared by Ministries of Finance, apparently without much coordination, despite the treaties’ broad policy impact. They may go unnoticed by the ministries and agencies that are not immediately involved.

Preferential trade agreements tend to be approached from a different angle. Their negotiation generally involves the same ministries and agencies that also coordinate their countries’ Doha Round offers. Thus, not surprisingly, the ensuing schedules show similar sector patterns: in particular, producer-related (intermediate) services tend to prevail over consumer services, apart from tourism. The fact that many PTAs are far more liberal in substance than current Doha Round offers may be attributed, inter alia, to the stronger (foreign-)policy connotations involved in bilateral or regional negotiations as well as governments’ interest, vis-à-vis other WTO partners, to preserve coinage for the final stages of the Round. There may be a perception that, under prevailing conditions, ambitious offers in the WTO could increase the appetite of potential PTA partners. While the outcome of the Doha Round, in terms of timing and substance, thus appears to be open at present, its sector profile is not difficult to predict. Don’t bet on healthcare.

  • 1 Annex C of the Ministerial Declaration provides that the main negotiating approach, via bilateral exchanges of requests and offers between interested delegations, be complemented by a plurilateral variant where like-minded Members form groups to pursue joint interests vis-à-vis others.
  • 2 Interestingly, Article XXVIII(l) defines a juridical person to mean ‘any legal entity duly constituted or otherwise organized under applicable law, whether for profit or otherwise’. Thus, references to profits exist in the Agreement, but not in the current context. See also R. Adlung: Public Services and the GATS, in: Journal of International Economic Law, Vol. 9, No. 2, pp. 455-485.
  • 3 A. VanDuzer: Health, Education and Social Services in Canada: The Impact of the GATS, in: J. M. Curtis, D. Ciuriak (eds.): Trade Policy Research 2004, Minister of Public Works and Government Services Canada 2004, pp. 388 and 395.
  • 4 Ibid., p. 404.
  • 5 In a similar vein: D.P. Fidler: Legal Review of the General Agreement on Trade in Services (GATS) from a Health Policy Perspective, Geneva 2004, WHO – Globalization, Trade and Health Working Paper Series, p. 36; and A. VanDuzer, op. cit., p. 447.
  • 6 For more details see R. Adlung, M. Roy: Turning Hills into Mountains? Current Commitments Under the General Agreement on Trade in Services and Prospects for Change, in: Journal of World Trade, Vol. 39, No. 6, 2005, pp. 1161-1194.
  • 7 J. Magdeleine, A. Maurer: Measuring GATS Mode 4 Trade Flows, in: WTO Staff Working Paper ERSD-2008-05, Geneva 2008, p. 18.
  • 8 F. Eschenbach, B. Hoekman: Services policies in transition economies: on the EU and WTO as commitment mechanisms, in: World Trade Review, Vol. 5, No. 3, 2006, p. 417.
  • 9 See, for example, V.G. Hatzopoulos: Killing National Health and Insurance Systems but Healing Patients?, in: Common Market Law Review, Vol. 39, No. 4, 2002, pp. 683-729.
  • 10 Document GATS/SC/32 of 15 April 1994.
  • 11 WTO document S/C/W/273 of 9 October 2006. The current version of the Guidelines is contained in WTO document S/L/94 of 28 March 2001.
  • 12 See also P. Sauvé, N. Ward: The EC-CARIFORUM Economic Partnership Agreement: Assessing the Outcome on Services and Investment, Brussels 2009, ECIPE, p. 35.
  • 13 A. VanDuzer, op. cit., p. 464 f.
  • 14 B. Gootiiz, A. Mattoo: Services in Doha – What’s on the Table?, in: Journal of World Trade, Vol. 543, No. 5, 2009, pp. 1013-1015.
  • 15 Situation as of September 2007. See J.A. Marchetti: Financial Services Liberalization in the WTO and PTAs, in: J.A. Marchetti, M. Roy (eds.): Opening Markets for Trade in Services: Countries and Sectors in Bilateral and WTO Negotiations, Cambridge 2008, Cambridge University Press, p. 327.
  • 16 Of the WTO disputes, only three dealt exclusively with GATS-related infringements (Mexico – Telecoms, US – Gambling, and China – Publications and Audiovisual Products). By the end of 2008, UNCTAD (IIA Monitor No. 1, 2009) counted 317 arbitration cases lodged under BITs. Of these, very few dated from before 1995. Using information from preceding UNCTAD studies, it may be assumed that about two-fifths of all claims concern Services.
  • 17 J.A. Marchetti, M. Roy: Services liberalization in the WTO and in PTAs, in: J.A. Marchetti, M. Roy (eds.), op. cit., p. 81 f.
  • 18 R. Adlung, M. Molinuevo: Bilateralism in Services Trade: Is there Fire behind the (BIT-) Smoke?, in: Journal of International Economic Law, Vol. 11, No. 2, 2008, pp. 40-43.
  • 19 See B. Choudhury: Recapturing Public Power: Is Investment Arbitration’s Engagement of the Public Interest Contributing to the Democratic Deficit?, in: Vanderbilt Journal of Transnational Law, Vol. 41, No. 3, 2008, pp. 792-797.
  • 20 P. Kulkarni: Impact of the GATS on Basic Social Services Redux, in: Journal of World Trade, Vol. 43, No. 2, 2009, p. 252.
  • 21 B. Choudhury, op. cit.
  • 22 Investment Arbitration Reporter, Vol. 2, No. 3 (10 February 2009) and No. 16 (14 October 2009).
  • 23 C. Blouin: Economic Dimensions and Impact Assessment of GATS to Promote and Protect Health, in: C. Blouin, N. Drager, R. Smith (eds.): International Trade in Health Services and the GATS, Washington DC 2006, The World Bank, p. 191.
  • 24 B. Choudhury, op. cit.
  • 25 I. Van Damme: Eighth Annual WTO Conference: An Overview, in: Journal of International Economic Law, Vol. 12, No. 1, 2009, p. 176.

Rudolf Adlung, Trade in Services Division, WTO Secretariat, Geneva, Switzerland.

Many thanks are due to Marie Sudreau for dedicated research assistance, and Antonia Carzaniga, Barnali Choudhury, Mireille Cossy and Nicolás Perrone for inspiring comments on an earlier draft. The usual caveats apply.


DOI: 10.1007/s10272-010-0341-8