Melbourne Journal of International Law
AN APPLICATION OF THE GATS DIGITAL TRADE FRAMEWORK
Arvin Kristopher Razon*
Blockchain technology and the World Trade Organization share the same principles of liberalisation and democratisation of trade. Relying on cross-border data flows and an open environment, blockchain platforms, such as Bitcoin and smart contracts, dispense with a central intermediary and distribute the function of validating transactions to a peer-to-peer network of participants, who maintain and update a record of all transactions in a transparent and immutable distributed ledger. The General Agreement on Trade in Services (‘GATS’) — as the only multilateral treaty that creates binding international obligations on the treatment to be accorded to service suppliers, including online services — provides a solid foundational framework for services supplied by participants on blockchain, even if, unsurprisingly, several grey areas exist. Applying GATS to blockchain means that many prohibitive regulations imposed by countries — from banning mining to criminalising the possession of cryptocurrencies — may be considered barriers to trade.
At its core, blockchain technology or distributed ledger technology (‘blockchain’) is defined by the same principles that underpin the World Trade Organization (‘WTO’): liberalisation and democratisation of trade. As the Fourth Industrial Revolution ushers in an era that harnesses modern technologies for the greater connectivity of people around the world, blockchain presents an opportunity to elevate international trade in services. At its core, blockchain is designed to democratise the supply of services through a primarily decentralised business model, eschewing the centralised supply of services that defined previous industrial revolutions.
With its capacity to enable real-time monitoring of risks and processing of transactions via a synchronised network of participants, blockchain can lead to an estimated USD1 trillion worth of savings in international trade, cutting physical transportation expenses of transacting parties by 20% and, by increasing efficiency in international supply chains, expanding global trade by around 15%. However, the path to the realisation of blockchain’s potential is not linear. Its uncertain privacy and cybersecurity implications, arguably due to a lack of a binding international legal framework, have resulted in diverging regulatory approaches.
This article argues that the 1994 General Agreement on Trade in Services (‘GATS’) may provide blockchain with a foundational legal framework that will facilitate its growth as an enabling technology. Applying GATS to blockchain is a challenge that likewise exists for other internet-based services, but GATS is flexible enough to address many relevant issues to blockchain and has the potential to expand global trade in services provided on a blockchain platform.
This article begins, in Part II, by describing the fundamental features of blockchain, characterised in this article as open and permissionless (as opposed to permissioned or controlled blockchain platforms), then proceeds to explain how two applications of this technology, cryptocurrencies and smart contracts, work. Bitcoin and Ethereum are two of the largest cryptocurrencies and smart contracts, respectively, in the global market. The varying governmental responses and blockchain’s reliance on cross-border data flows are important in framing blockchain as a digital trade issue. Part III applies GATS as a legal framework to fill the legal vacuum in which blockchain operates. This part also clarifies whether blockchain, particularly in relation to cryptocurrencies and smart contracts, should be classified as a good or service, and whether it can fall squarely under any of the four modes of supply in GATS. Part IV discusses the service sectors relevant to blockchain, as well as an application of the core GATS disciplines to blockchain. Part V gives an overview of the ‘safety valves’ that WTO members may use to deviate from their applicable commitments under GATS. Lastly, Part VI explores parallel and complementary developments that can help secure the liberalisation of blockchain.
This part discusses the nature of blockchain, the varying governmental responses to it and its reliance on cross-border data flows. Together, these factors illustrate how blockchain is a digital trade issue.
The beginnings of blockchain can be traced to Bitcoin, which uses this technology as its digital infrastructure. In 2008, a person under the pseudonym Satoshi Nakamoto released a white paper on Bitcoin, a blockchain-based cryptocurrency. He conceptualised and launched an electronic cash system that not only facilitates transactions but also generates its own currency, without the need of financial institutions or a government central bank. Dispensing with a third party, Bitcoin uses a digital infrastructure consisting of an immutable distributed ledger that spreads the centralised functions of an intermediary to participants in a peer-to-peer network. The infrastructure has no single point of system failure, as the ledger is maintained not by one intermediary, but by all participants. The ledger is a chain that contains a history of all transactions ever performed in the network and is publicly available to all participants. Participants within the network validate the transactions. Simultaneously-occurring transactions are then hashed and formed into a block, which is then immutably and irreversibly appended to the chain.
Since the introduction of Bitcoin, various uses have been conceived for blockchain as an underpinning architecture, with smart contracts as another leading application. Ethereum, currently the most popular platform for smart contracts, allows the automated execution of smart contracts by embedding computer code on top of the blockchain platform: computer language is used to design and implement parties’ agreements, instead of legal language in agreements drafted by lawyers and implemented by various parties. Whereas Bitcoin participants validate the transactions in order for the transaction to push through, Ethereum participants use blockchain’s peer-to-peer network feature to collectively validate certain conditions that trigger the automated execution of the agreement. With automation, blockchain is a mechanism to guarantee the performance of an agreement, minimising the tendency of parties to alter their obligations after its perfection.
Some actions that may be automatically executed through smart contracts include the calculation of interest, transfer of payment for goods and storage of records. Ethereum also allows the automation of complex, multi-party agreements: loan financing with the buyer, seller and lender all having a stake in the platform; distribution by a company of its shares and dividends; and even the provision of public services to constituents. At this juncture, whilst Bitcoin is arguably a form of smart contract, with a predetermined purpose specific to the transfer of cryptocurrencies, this article refers to smart contracts in the general sense, that is, as a flexible platform that allows parties to design various agreements. Further, the GATS analysis in this article generally pertains to the fundamental features of blockchain as a foundational technology, and specifically refers to smart contracts and cryptocurrencies, where important distinctions and nuances may vary the analysis.
Distilled to its essence, blockchain pertains to the transparent, secure and immutable distributed ledger, powered by a decentralised peer-to-peer network of participants. Its value proposition lies in the strength of its ledger, secured by hashing, cryptography and asymmetric key encryption. The unique functioning of blockchain opens a number of possibilities for global trade. Its infrastructure has a self-sustaining capacity to validate and verify transactions, record them in an inalterable state and subsequently broadcast the history of transactions. It is an unprecedented attempt to democratise traditionally centralised functions, possibly leading to the economic development of its participants, who may be both consumers and also service suppliers with decision-making and voting capacity. With blockchain, the need for a central governing source is done away with, as consumers and service suppliers may be one and the same. Because there is no central intermediary, open, permissionless blockchain platforms (which users can join without the need of an approving entity) require a critical mass of users to succeed and participate, so national borders need to be as open and unrestrained as possible.
The Organisation for Economic Co-operation and Development (‘OECD’) has recognised the impact of technology on economic growth and social welfare, and how governmental responses to novel challenges posed by emerging technologies can influence their trajectory. Governmental approaches to blockchain can be organised into three categories: (a) liberalised or laissez-faire, (b) prohibitory and (c) prudential. For some countries, an unstable regulatory approach has led to businesses establishing their operations in other countries. To be clear, countries with a prohibitory approach have issued regulations banning cryptocurrencies, but not blockchain in general. On the other hand, countries with a liberalised or prudential approach have issued regulations relating to blockchain in general, typically recognising that cryptocurrency is one of its most popular uses.
Under the first category are countries that are generally receptive to blockchain: the United States Securities and Exchange Commission’s approval of share issuance via blockchain, Germany’s blockchain platform for trading digital assets, South Korea’s generally ‘hands off’ approach and the United Kingdom’s Office of Science report on the benefits of blockchain for public governance. Mining, the process of validating transactions and creating blocks in the chain (as explained in Part IV(A)(3) below), is expressly allowed in Belarus and is considered a legal entrepreneurial venture in Ukraine.
Countries that have expressly banned cryptocurrencies (but not blockchain) have adopted a prohibitory approach. In Bangladesh, individual Bitcoin users could be imprisoned. Vietnam has issued a warning that using cryptocurrencies may result in prosecution, even seizing BitcoinVN, the country’s largest cryptocurrency exchange. Ecuador decided to ban cryptocurrencies, only to create its own government-backed digital currency. Algeria considers mere possession of Bitcoin a crime.
Countries with a prudential approach have chosen to tighten their blockchain-related regulations, specifically in relation to cryptocurrencies, without imposing a total ban. Gibraltar has issued comprehensive regulations on consumer protection rules for blockchain. Finland has chosen to focus on taxation, by considering Bitcoin to be earned income from a private transaction. The Philippines allows licensed cryptocurrency exchanges to operate, subject to reporting requirements.
China is an interesting case study, as it has gone from having one of the largest Bitcoin exchanges, to a plan to ban cryptocurrency mining altogether, while the president publicly endorses blockchain as a technology. As China accounts for around 70% of Bitcoin mining, the People’s Bank of China and Administration of Foreign Exchange wants to ban Bicoin mining and has already blocked access to Bitcoin exchange services and ordered mining companies to make an ‘orderly exit’. China’s regulatory approach is already shaping the direction of the industry, with the companies initially choosing China for its cheap electricity and labour, but now moving to Singapore, the US and Canada.
Against the backdrop of these widely divergent regulations is a call for the G20 member countries to address the divergent approaches on blockchain. In expressing support for blockchain development, the G20 recognises the unique capacity of blockchain to create a decentralised economic order, where people directly transact with each other in the global market. The possible gains from blockchain are clear. For blockchain users, transactions are made more secure with a robust and active network of participants. For developing countries, blockchain is a possible engine for financial inclusion to billions of unbanked people. It can generate jobs and tax revenues. Most importantly, prohibiting blockchain at its early phase will stifle innovation and prevent experimentation that may lead to better applications of this technology: countries that have adopted the wait-and-see approach have seen more innovation that continues to push the boundaries of blockchain. A liberalised approach tempered by appropriate safeguards, deliberately adopted by the international community, appears to be a viable solution that will allow participants from anywhere in the world to benefit from blockchain.
Before proceeding to apply GATS, a determination of whether blockchain involves a supply of service — and not goods — is relevant. How GATS may apply to ‘new’ services that did not exist at the time of its adoption, such as blockchain, is also important to discuss, as this will show the relevance of GATS in today’s digital economy.
Characterising blockchain either as a good or a service is important in determining whether the General Agreement on Tariffs and Trade (‘GATT’) or GATS would apply. While GATS applies to trade in services, GATT applies to trade in goods. If the goal was to establish full liberalisation of trade, GATT would be preferable: as it follows a negative-list approach, obligations under GATT apply by default and may be qualified only by narrow exceptions. On the other hand, GATS follows a positive-list approach, which means some of the most important obligations only apply if members expressly choose that they do so. The line between goods and services is not clearly demarcated, as many products possess the characteristics of both.
Are blockchain applications, therefore, goods or services? The inherently digital nature of blockchain is an important factor in characterising blockchain as more of a service than a good. The likeness or similarity of blockchain with existing goods or services already covered by GATT or GATS will determine the applicability of either instrument. The criteria for ‘like or similar products’ under GATT varies on a case-by-case basis, although certain criteria are generally agreed upon for determining whether a product is similar: physical likeness and interchangeability, end uses, consumers’ tastes and habits, and substitutability. The tangibility of goods is an important factor, although its exact role remains to be decisively settled, as the question of whether GATT applies to digital products traded electronically continues to be a point of disagreement among members. With this, electronic products, such as downloadable media that used to be delivered in a physical form, cannot be definitively subsumed under the concept of ‘like products’ under GATT.
GATS, on the other hand, has a more promising case in providing a better foundation for covering blockchain technologies. Digital transactions do not generally include tangible products, and the sale of electronic products often requires the continual delivery of services — whether for support or for maintenance — for them to be even usable. Many electronic products are covered by GATS in that their advertising, payment, delivery and distribution all significantly involve an electronic service. Moreover, blockchain essentially involves the validation by users of transactions conducted over the platform. Whilst users cannot do so without their computing devices, blockchain still intuitively involves the supply of a service and not the sale of a useful electronic product.
Whilst digital services are likelier to be classified as services, rather than goods, applying GATS is hardly a straightforward exercise. Two specific features of blockchain pose a challenge to the application of GATS: the fact that blockchain is a service performed on the internet (and not a physically rendered service), and its relative nascence (as a service that arguably did not exist and had no physical counterpart at the time GATS was adopted). Both issues are addressed in the next section.
Whilst GATS — as well as GATT and other WTO agreements — has not been amended to react to digital innovation and the internet, the relevant cases emphasising technological neutrality and its underpinning principles highlight the flexibility and resilience of the most important GATS obligations to internet-related services — as effectively as, or perhaps even better than, made-to-measure international regulations can.
First, the digital nature of services being delivered is not necessarily an obstacle to the application of GATS. In the Work Programme on Electronic Commerce adopted by the WTO Council for Trade in Services (‘CTS’), electronic delivery of services was generally viewed as falling within the scope of GATS. The technological neutrality principle — that no distinction should be made between offline and online delivery of services — has been an important factor in WTO decisions applying GATS to digitised services. Although countries continue to disagree on the extent to which GATS applies to digital services, the WTO Panel in United States — Measures Affecting the Cross-Border Supply of Gambling and Betting Services (‘US — Gambling’) expressed the view that GATS rules are applicable to services provided in the digital marketplace, noting that this principle appears to be ‘largely shared among WTO Members’. Later, in China — Measures Affecting Trading Rights and Distribution Services for Certain Publications and Audiovisual Entertainment Products, the WTO Appellate Body found that a commitment on sound recording distribution services extended to sound recordings distributed through the internet. Both of these case law developments affirm the evolutionary nature of GATS.
Secondly, the question turns on how GATS can apply to ‘new’ services that arguably did not exist at the time of its adoption. The concern stems from the inappropriateness and outdatedness of the services classification under GATS, where members make scheduled commitments based on services listed in the GATS Services Sectoral Classification List (‘GATS Classification List’), which then refers to the 1991 Provisional Central Products Classification by the United Nations (‘CPC (1991)’). What complicates the analysis is that during the Uruguay Round, ‘computer and related services’ as a classification was not nearly as relevant then as it is now and does not reflect the complex reality of digital services. Indeed, how can WTO members include a service that they did not at the time even imagine was possible? Thus, members then did not have a problem with making far-reaching national treatment and market access commitments — leading to a relatively high level of liberalisation amongst members, and consequently a relatively tight leeway for them to impose prohibitive domestic regulations in computer and related services under GATS. Many of these ‘new’ services are ‘over-the-top’ services, which deliver content using an internet platform without the involvement of the telecommunications provider. As described, blockchain could be considered as an over-the-top service, as the validation of transactions happens over the internet.
The apprehension regarding the treatment of ‘new’ services stems more from the limitations of the service classification under the GATS Classification List and CPC (1991) than from the merits of GATS as a legally binding instrument. However, classifying ‘new’ services and fitting them into existing commitments of WTO members is not impossible under GATS. Many ‘new’ services could still be accommodated by the broad definitions under both the GATS Classification List and the CPC (1991). Also, whilst the GATS Classification List or the CPC (1991) (drafted back in 1991 and 1990, respectively) may be outdated to provide guidance in classifying new services, the new version of the CPC (1991), the latest of which came out in 2015 (‘CPC (2015)’), refers to internet and digital services, and may provide better guidance for members. The WTO Appellate Body has also expressed that ‘although Members were encouraged to follow the broad structure of W/120 [GATS Classification List], it was never meant to bind Members to the CPC definitions’. Lastly, in a case where a member’s commitment to a particular service does not refer to CPC codes, the ordinary meaning, objective and purpose of that service classification may provide guidance and do not necessarily foreclose digital services.
The relevance of GATS to blockchain and other ‘new’ digital services also resonates with broader policy considerations underlying many international framework agreements. In this regard, it is important to emphasise that GATS is a multilateral agreement (not an agreement among private parties) and is meant to be a long-term arrangement (not an ad hoc transaction). The flexibility of its provisions mirrors the intent to make it adaptive to rapid technological changes. In light of such policy considerations, GATS must be viewed as a dynamic process and must be interpreted according to reality and modern progress, with the mindset of pursuing its purposes (the liberalisation of trade in services) and not undermining its very existence in the long run. Hence, whilst GATS may be lacking, WTO dispute settlement bodies are unlikely to interpret GATS as wholly inadequate to address digital trade issues, as such a limited approach may render it obsolete in modern trade.
Blockchain is clearly a ‘new’ digital service that did not exist at the time GATS was adopted — along with video streaming, cloud storage and telecommuting. But as explained above, its nascence does not mean that GATS is wholly irrelevant, because of the WTO’s technologically neutral approach and the wide latitude that the interpretation of members’ existing commitments allows.
How could specific blockchain applications be framed as services under GATS? With regard to cryptocurrencies, Bitcoins are not merely virtual equivalents of fiat money (and therefore cannot be considered goods in the same way that money stored in virtual accounts can be considered as such). Rather, Bitcoin is a virtual currency with no underlying asset, not backed by any central bank and existing only through the services of the network participants within the platform. Bitcoins are generated as a reward for participants who support the network, ‘analogous to gold miners expending resources to add gold to circulation’. In other words, Bitcoins motivate participants to supply their services, that is, the verification and validation of transactions in the ledger. The ‘supply of a service’ under GATS art XXVIII(b) — defined as the ‘production, distribution, marketing, sale and delivery of a service’ — approximates the services supplied by participants in a Bitcoin network.
Services provided under smart contracts likewise fit the ‘supply of a service’ definition under GATS. Developers who code the smart contracts onto the platform provide a service by writing the code in the same way that lawyers draft contracts. For example, a smart contract can be used by various parties in a global supply chain to document the trade process, enrol and approve relevant documents (such as letters of credit and bills of lading), manage the shipment process and release the goods to the buyer — all specific points that depend on the authentication of participants on the blockchain. As the goods move throughout the supply chain, participants will create blocks that validate the movement of goods, triggering the release of payment or title to the goods to the proper party. In return for their validation services, participants are compensated (in Ethereum, through the payment of ‘Ether’).
Is GATS the appropriate vehicle for the liberalisation of blockchain? Aside from providing a legal foundation by meeting the criteria of how GATS can apply, another policy reason to apply GATS is that its fairly comprehensive rules and commitments provide the stability needed for the blockchain to grow, with the participation of its many members and with as few protectionist measures as possible. As a technology reliant on cross-border data flows, blockchain can only prosper with a transparent and liberalised multilateral framework for trade, removing trade barriers that may get in the way of economic growth and development, as set out in the GATS Preamble.
Under GATS art I:3(a), measures by members may be taken by central, regional or local governments, and even by non-governmental bodies in the exercise of delegated governmental authority. Such measures may in substance relate to the purchase, payment or use of services, or may pertain to access and use of public ancillary services necessary for the delivery of services. Examples include the imposition of burdensome administrative requirements to the service supplier, or a complete ban on domestic customers from using such foreign services.
The varying governmental responses to blockchain, as previously discussed in Part II(B), may be characterised as measures covered by GATS, particularly the prohibitory measures that filter access to blockchain apps or services (therefore preventing service suppliers from offering their services to countries that prohibit blockchain), or even criminalise ownership of cryptocurrencies. Such measures, in so far as they constitute barriers to international trade in services and violate existing commitments of a member, may be challenged before WTO’s dispute settlement system, dubbed the ‘bedrock of the multilateral trading system’ and the ‘jewel in the crown of the WTO’.
In addition to classifying e-commerce as a service covered by GATS (as discussed in the immediately preceding section), the WTO has had the occasion to discuss digital services in the context of the relevant modes of supply under art I:2: (1) cross-border supply, (2) consumption abroad, (3) commercial presence and (4) presence of natural persons. Specifying which of the four modes of supply applies is relevant because members’ commitments differ depending on the services concerned and the corresponding mode of delivery.
In the supply of digital services, either Mode 1 (cross-border supply) or Mode 2 (consumption abroad) is relevant. In determining which mode of supply applies to digital services, two questions may be asked: does the service supplier virtually travel to the consumer’s country to provide the service (Mode 1) or does the consumer travel cross-border to the service supplier’s country (Mode 2)? Generally, commitments under Mode 2 are more liberal than under Mode 1. Also, Mode 2 triggers the application of the domestic laws of the service supplier, while Mode 1 triggers the application of the consumer’s domestic laws. Regulatory measures in the applicable jurisdiction would ultimately apply if GATS commitments of that member have been violated.
This question has not been settled. In US — Gambling, the Panel held that internet-based services could fall under Mode 1, first by characterising the service (in this case, an online gambling service provided by Antiguan service suppliers) as involving ‘remote supply’, then concluding that this pertains to cross-border supply when the supply is between territories of different members. The Appellate Body affirmed this view and analysed the US’ Mode 1 commitments. This finding avoids controversy: construing digital services under a Mode 2 supply would displace the consumer’s local laws and apply the domestic laws of the foreign service supplier, without the usual notice of entry into a foreign jurisdiction accompanying foreign travel.
At this point, to further understand the services provided in a blockchain, users and participants in a blockchain must be defined. Users conduct transactions in blockchain, and may be considered consumers: in Bitcoin, they may refer to transacting parties that transfer Bitcoin to other parties; in smart contracts, they may be transacting parties to the contract who require the distributed ledger to automate the execution of their respective obligations. Participants, on the other hand, are units that use their computing systems to create new blocks to be appended on the chain. They are, in essence, service suppliers in the GATS context. An important feature of blockchain is that users may participate in the validation of transactions and creation of blocks by using their own devices, or by sharing computer resources with other users — provided that they are not part of the transaction being validated. In other words, users (consumers) may be participants (service suppliers) for other transactions, but not for a transaction that involves them.
The implication of a Mode 1 supply in blockchain is that a network participant, as a service supplier, virtually travels to the consumer’s territory, and is subject to the latter’s jurisdiction. In blockchain, the consumers are the transacting users, ie the parties involved in the transfer of payments (for Bitcoin) or the parties in a smart contract (for Ethereum). They need the peer-to-peer network of participants for their transactions to go through. The network participants then supply their services by validating the Bitcoin transactions or the conditions of the smart contract. Construing blockchain-based services under a Mode 1 supply may be unnecessarily complicated, if not infeasible, as network participants would be simultaneously travelling to the locations of transacting users (who are likely located in various jurisdictions, and there would, therefore, be no clear applicable law), or the object of a smart contract may not even be within the jurisdiction of contracting parties altogether (in which case, again, there would be no clear applicable law, as the consumers’ locations would not be material).
On the other hand, construing services under a Mode 2 supply would simplify the application of GATS. In this case, the applicable law would unequivocally be that of the service supplier (the network participant). Whilst the WTO has not expressly considered this, a Mode 2 supply would mean that the transacting users, by asking the peer-to-peer network of participants in a blockchain to validate their transaction, virtually travel to the location of each network participant to request a validation of their transaction. In this case, the domestic law of the service supplier concerned would apply.
Aside from the broad definition of services under GATS, the provision of services is generally susceptible to various interpretations by researchers, governments and trade negotiators, depending on their political interests. For online services like blockchain, the enduring question is whether they form a ‘new’ service or are simply a development of an existing service covered by a member’s schedule.
Part III(B) explains why GATS still applies even if blockchain is a ‘new’ service, because the limitations of the GATS Classification List and the CPC can be overcome in other ways, and because of broader policy considerations. GATS case law also supports this position. In China — Publications and Audiovisual Products, China attempted to argue that sound recordings distributed on the internet did not exist at the time of its accession, and hence could not be accommodated by its scheduled services. Both the Panel and Appellate Body rejected this argument, with the Appellate Body ruling that interpreting GATS commitments based on its meaning at the time China’s Schedule was concluded would mean that service sectors would have different meanings, content and coverage based on a member’s date of adoption. This would ultimately ‘undermine the predictability, security and clarity of GATS specific commitments’.
The Appellate Body in particular examined whether ‘sound recording’ and ‘distribution’ are couched in generic enough terms that what they apply to may broaden over time. In ruling in the affirmative, the Appellate Body argued that its reading of China’s commitment in its GATS Schedule aligns with its ruling in United States — Import Prohibition of Certain Shrimp and Shrimp Products, where ‘exhaustible natural resources’ must be ‘read by a treaty interpreter in the light of contemporary concerns of the community of nations about the protection and conservation of the environment’. The Appellate Body believed that ‘natural resources’ must be interpreted not as a static concept, but in an evolutionary manner. Applying this reasoning, the Appellate Body concluded in China — Publications and Audiovisual Products that based on the ‘objective and purpose’ analysis, sound recording distribution services can logically extend to the electronic distribution of sound recordings, which did not exist at the time GATS was adopted.
Whilst blockchain may not have a physical counterpart in the same manner electronic sound recordings do, the reasoning in China — Publications and Audiovisual Products is nonetheless applicable. If existing commitments and services classifications are generic enough that what they apply to may broaden over time — specifically to accommodate blockchain technologies — then perhaps GATS is not too obsolete to deal with blockchain after all. Fitting blockchain within existing obligations would result in stability and predictability.
Besides, revising the schedules is a daunting undertaking that will take several years and protracted negotiations. This part explores the idea of fitting blockchain within the existing services classification, and the applicable obligations under GATS.
Under GATS art XX, a member must submit a schedule of commitments, which forms an integral part of GATS itself. As GATS follows a positive-list approach, its most important disciplines, the market access and national treatment obligations, only apply where a member has undertaken these commitments to service sectors listed in its schedule. Service sectors that do not fall within the sectoral classification are deemed not liberalised. If not expressly listed, a service must be construed as somehow included in a service sector listed in a member’s schedule in order for GATS commitments to apply. Thus, the main question to be asked under this section is: what is the primary function of blockchain-based services?
When scheduling their commitments, members have a broad discretion in describing the services to which they are committing. While not obliged, a member is encouraged to use the GATS Classification List, the CPC (1991) or a similarly precise classification system. The sub-sectors in the GATS Classification List are aggregated categories of the more comprehensive list in the CPC (1991), and often make explicit references to their corresponding codes in the CPC (1991), although the two are considered different. Moreover, the GATS Classification List does not explain the scope of each service activity, so the CPC (1991), as the source document, may be referred to. In practice, a vast number of existing schedules of members follow the GATS Classification List, including its structure and headings, and often refer to CPC codes in the CPC (1991) to describe their commitments. In case the CPC has not made the basis for a member’s commitment, the succeeding CPC versions or the ordinary meaning of the services may provide meaningful guidance. The latest CPC, CPC (2015), contains many updates regarding digital services that may be helpful in analysing the services classification of blockchain.
Blockchain potentially falls under the following four types of services under the GATS Classification List, with its corresponding CPC code in the CPC (1991) reproduced below. The analysis below explores the idea that classifying blockchain is possible by working within the existing GATS structure and members’ current commitments.
GATS Classification List
1 Online information and database retrieval
Sub-sector 2.C.j (under Telecommunications Services, sub-sector of Communication Services sector)
7523 (data and message transmission services)
2 Database services
Sub-sector 1.B.d (under Computer and Related Services, sub-sector of Business Services)
844 (database services)
3 Online information and/or data processing (including transaction processing)
Sub-sector 2.C.n (under Telecommunications Services, sub-sector of Communication Services)
843 (data processing services)
4 Data processing services
Sub-sector 1.B.c (under Computer and Related Services, sub-sector of Business Services)
843 (data processing services)
Online information and database retrieval refers to CPC (1991) class 7523. As the CPC (1991) does not provide a definition of this general class of data and message transmission services, its subclasses may be helpful:
7523 Data and message transmission services
75231 Data network services
Network services necessary to transmit data between equipment using the same or different protocols. This service can be provided via a public or dedicated data network (ie via a network dedicated to the customer’s use).
75232 Electronic message and information services
Network and related services (hardware and software) necessary to send and receive electronic messages (telegraph and telex/TWX services) and/or to access and manipulate information in databases (so-called value-added network services).
Based on an analysis of these subclasses, it can be inferred that CPC (1991) class 7523 only covers the underlying hardware and network services necessary for data transmission, and not the delivery of information online. While network services are important for participants to provide their services in a blockchain, in the sense that a stable and consistent connection to the internet is necessary, these are peripheral to the core services involved in blockchain: the validation and verification of transactions, and creation of blocks to be appended to the chain. Thus it does not seem that blockchain falls within this sub-sector.
CPC (1991) class 844 defines database services as ‘[a]ll services provided from primarily structured databases through a communication network’, excluding ‘[d]ata and message transmission services’ under class 7523. This involves two requirements: (a) that the services be provided from primarily structured databases and (b) that the means of provision is a communication network.
On the surface, blockchain may be regarded as a database, as its main ‘product’ is a ledger of organised data, consisting of blocks that contain transaction data, a hash that references the previous block, addresses of the recipients and senders and the digital signature of transacting parties. Participants provide their computational resources in order to ensure that each block contains a uniform and consistent pattern of data.
However, there is an important difference between a database and a blockchain ledger. Whilst a database refers to organised data meant to provide information, a blockchain ledger is meant to make transactions secure and confidential in an otherwise trustless environment. Users and participants in a blockchain platform do not consult the ledger to obtain information about transactions that have transpired in the same way that people consult a database to retrieve information. Rather, blockchain users transact with each other (ie transfer Bitcoin or create smart contracts to govern their arrangements) in a blockchain platform, knowing that such transactions are conducted in a secure environment, validated by participants in the network. In other words, a blockchain ledger is a function of blockchain’s transparency and immutability, and a by-product of services provided by blockchain participants, who are primarily concerned with the validation and verification of transactions. In conclusion, a database is primarily a storage of information, whereas a distributed ledger is a way to record and log movements in the blockchain.
An anomaly in the GATS Classification List is the reference of both ‘online information and/or data processing’ (under telecommunications services) and ‘data processing’ (under computer and related services) to the same class, class 843 (data processing services), under the CPC (1991). The CTS notes the overlap, and that ‘it may not be clear when telecommunications services, computer services, or both are being supplied’. Should blockchain be considered as data processing services under telecommunications services or computer and related services?
To resolve this apparent conflict in determining whether a service falls within the Telecommunications Service or another kind of service, a key distinction must be made between use and supply. When telecommunications infrastructure is used by service suppliers as a means of delivery for other services (such as financial services or video streaming), such suppliers are considered mere users of telecommunications networks and services, and therefore retain their original category (as financial services providers or video distribution providers). This distinction is consistent with the definition of telecommunications under the Annex on Telecommunications to GATS: ‘the transmission and reception of signals by any electromagnetic means’.
For example, a cloud computing platform retains its computer services classification, even as it uses telecommunications networks for its delivery of cloud computing services. A blockchain platform benefits from a similar analogy: whilst blockchain requires the use of telecommunications networks for the delivery of services within the network, such networks are merely ancillary to the main service delivered by participants in the platform. Therefore, blockchain platforms are not necessarily telecommunications service providers, as they only use telecommunications services but do not supply them.
Classifying blockchain as data processing services under computer and related services arguably makes sense. Data processing services as a general class does not have a definition. Definitions are provided in the CPC (1991) subclasses, as follows:
843 Data processing services
8431 84310 Input preparation services
Data recording services such as key punching, optical scanning or other methods for data entry.
8432 84320 Data-processing and tabulation services
Services such as data processing and tabulation services, computer calculating services, and rental services of computer time.
8433 84330 Time-sharing services
This seems to be the same type of services as 84320. Computer time only is bought; if it is bought from the customer’s premises, telecommunications services are also bought. Data processing or tabulation services may also be bought from a service bureau. In both cases the services might be time sharing processed. Thus, there is no clear distinction between 84320 and 84330.
8439 84390 Other data processing services
Services which manage the full operations of a customer’s facilities under contract: computer-room environmental quality control services; management services of in-place computer equipment combinations; and management services of computer work flows and distributions.
What stands out from the above definitions are the references to ‘input preparation’, ‘tabulation’ and ‘computer calculating’ services. These explanations illustrate data processing as a form of converting discrete data into a systematically organised and useful form.
Do Bitcoin transactions involve data processing in the manner contemplated under computer and related services? A more complete picture of services provided in a Bitcoin platform would tie it to data processing. In this regard, Bitcoin participants make three specific ‘actions’ in the Bitcoin platform: (a) mining, (b) appending the block onto the chain and (c) maintaining a synchronised copy of the ledger.
At any given time, senders transfer Bitcoins to recipients in a Bitcoin network. In a traditional payment system (eg bank transfers), an intermediary such as the bank will validate this transaction and facilitate the transfer of Bitcoins. However, in Bitcoin, the transaction cannot go through until the subsequent steps are followed:
1 First, every transaction in a Bitcoin network is broadcast to all participants (ie those who own computing systems and perform the data processing service under GATS).
2 Each participant forms a block comprising an average of 500 Bitcoin transactions.
3 The participants then race to solve a difficult cryptographical puzzle generated by the platform, in a process that is called mining. In this part, a computing system ‘mines’ for a solution to a randomly generated mathematical equation. The solution generated is called a ‘proof-of-work’, and a miner who correctly generates this first is incentivised with Bitcoins or transaction fees.
4 When a participant finds a proof-of-work, the block is broadcast to all the other participants in the network. The participants will only accept the block if the transactions in it are valid, and if the proof-of-work is correct; this is called a ‘consensus protocol mechanism’. Each participant is involved in this cross-checking process, although it expends less energy than mining.
5 Once a block is accepted, it is appended to the chain, which is replicated and synchronised across all participants in the network. The next block will be appended to this previously accepted block, using the hash of the previous block as the jumping point. Every participant will have a real-time replicated copy of the ledger.
Services provided by participants in smart contracts in Ethereum essentially follow the same process. Participants provide their computing resources to validate certain movements pre-coded into an agreement, and ‘blocks’ are appended to the chain to trigger the performance of certain obligations, such as the release of payment, title to goods or dividends to shareholders.
Thus data processing is inexorably connected to every action done by a participant on a blockchain platform. From creating a block and mining for proof-of-work to appending the block on a chain and synchronising the ledger across the network, the crux of a participant’s service is data processing: to engage in the tabulation and intensive calculation of data to form a block (class 8432) and then to record that block as an appendage into the ledger (class 8431), ultimately enabling transactions between users of the platform. Without participants’ data processing services, the performance of obligations under a smart contract or the transfer of Bitcoins cannot be perfected. The Panel’s decision in China — Certain Measures Affecting Electronic Payment Services is also relevant. In that case, the Panel interpreted a commitment regarding ‘all payment and money transmission services’ to mean an intention to comprehensively cover the entire spectrum of services under this sector. This would include ‘all services essential to payment and money transmission, all means of payment and money transmission ... and all associated business models’. Applying this case, all services — mining, appending of blocks onto the chain, and maintenance of the ledger — provided by a blockchain participant are subsumed under data processing services.
Again, the analysis above is useful in determining a member’s commitment to a service under the GATS Classification List, insofar as it drafted its schedule using the CPC (1991). In case the CPC (1991) was not used by a member, CPC (2015) could be considered as a reference for GATS definitions. In this case, CPC (1991) class 84290 (other internet telecommunications services) is a possible classification for blockchain-related services.
Further, although the analysis above is grounded on a practical understanding of how blockchain works, blockchain is still a new technology that did not exist at the time GATS was implemented. Thus, an entirely new service classification that expressly includes blockchain and upon which members could expand their commitments will in the end still provide the legal certainty that the status quo cannot guarantee.
With its goals of assuring equal competition for both domestic and foreign service suppliers, liberalising markets and decreasing protectionism through a rule-based system of fair trade, the WTO is underpinned by free trade principles, such as national treatment, most-favoured nation (‘MFN’) and market access obligations. MFN and national treatment disciplines mainly refer to discriminatory measures, while market access refers to quantitative measures that may either be discriminatory or non-discriminatory. These principles are important in sustaining the growth of blockchain from its early stages of development.
Under GATS art II:1, each member must ‘accord immediately and unconditionally to services and service suppliers of any other Member treatment no less favourable than that it accords to like services and service suppliers of any other country’. Under this principle, a member must give services and service suppliers from other members equal treatment it accords to like imported services and suppliers from another country, whether or not they are a WTO member. This discipline therefore assumes that a member has accorded certain trade benefits to other countries, and all other members must benefit from the same.
Equal treatment depends on whether service and service suppliers are in the first place alike. The ‘likeness’ standard is not defined under GATT or GATS, although it has been interpreted under GATT on a case-by-case basis, taking into account, among other factors, the international classification for tariff purposes, ‘end uses, physical characteristics and consumer habits’ — all of which, to the extent applicable, may be relevant to GATS. Transposing this test to GATS may be useful to a limited extent, with end uses and consumer habits as the most relevant.
Considering that blockchain is an enabling technology, with many possible applications such as cryptocurrencies, smart contracts, identity management and recordkeeping, the scope of ‘likeness’ must be defined for purposes of applying MFN obligation. WTO dispute resolution bodies may apply a broad standard, where various blockchain applications, insofar as they share the same core technology and infrastructure, may be considered ‘like services’. The criteria for ‘likeness’ here that may arguably be used is the similarity of characteristics of blockchain applications. Under this idea, a member cannot, without running afoul of its MFN obligation, on one hand allow supply chain management solutions using blockchain from a service supplier in another member, and then on the other hand prohibit cryptocurrencies offered by a service supplier in another member. However, strictly applying the ‘likeness’ test would likely mean that a member cannot, for example, allow cryptocurrencies from one country and disallow cryptocurrencies from another member — as cryptocurrencies more or less operate within the same market.
To complement the MFN obligation is the national treatment obligation, which ensures that ‘like’ foreign services or service suppliers provided by other members may not be treated less favourably in comparison to domestic services or service suppliers. Unlike MFN, the scope of a member’s commitment to national treatment depends on its schedule.
To assess a possible violation of this obligation, the domestic and foreign service suppliers must be operating in the same market, and a general discriminating measure that could modify the competitive conditions in the market must exist. Also, the scope of the ‘no less favourable treatment’ permits formally identical or formally different treatment, as long as it does not modify the conditions of competition in favour of domestic services or service suppliers compared to like services or service suppliers of any other member. Thus, a member may impose different formal registration requirements for blockchain service suppliers without violating its national treatment obligation. However, where the measure provides cost advantages, such as subsidies or tax breaks, that favour domestic blockchain start-ups — with the goal of creating a domestic blockchain start-up ecosystem at the expense of foreign blockchain service suppliers — this obligation may be violated. In assessing a violation of the national treatment obligation, the effect of the measure must be assessed depending on its impact on competitive conditions.
Also, as in MFN, discrimination is based on whether foreign services or service suppliers are ‘like’ those provided by domestic services or service suppliers. Measures where a country prohibits blockchain but then proceeds to put up its own blockchain platform may violate national treatment. This brings to mind Ecuador, which banned Bitcoin in favour of creating its own national digital currency. In this regard, cryptocurrencies are similar in their end uses. The national treatment obligation may also be broadly interpreted by WTO tribunal bodies to extend to all the applications of blockchain. For example, if a country expressly prohibits cryptocurrencies from foreign service suppliers but proceeds to put up its own land registration platform powered by blockchain, the national treatment obligation may not be violated.
Considered as the basic instrument to liberalise international trade, market access of foreign members to service sectors and modes of supply in another member is not a default and may be subject to conditions, as specified in its schedule.
Under GATS art XVI:2, there are six specified restrictions pertaining to limitations on: (a) the number of service suppliers, (b) the total value of service transactions or assets, (c) the total number of service operations or total quantity of service output, (d) the total number of natural persons who may be employed in a particular sector, (e) the types of legal entity through which service may be supplied and (f) the participation of foreign capital. Article XVI defines market access through examples without providing a general definition for the term, which raises the question: what is the essence of the market access discipline?
In US — Gambling, the measures involved the US’ prohibition against providing ‘remote’ gambling and betting services, in violation of GATS arts XVI:2(a) and XVI:2(c). The Panel’s finding, affirmed by the Appellate Body, was that the total ban against providing these services was in effect a ‘zero quota’ broadly covered by market access obligations. Applying this reasoning to blockchain, a member’s total prohibition and even criminalisation of the use or possession of cryptocurrencies may violate market access obligations if commitments on data processing services under computer and related services have been undertaken.
Under GATS art III, transparency is a general obligation that requires members to publish promptly all relevant measures of general application pertaining to or affecting GATS. ‘Enquiry points’ to provide specific information on such measures must also be established by its members. A corollary to this is art VI, which requires members to apply domestic measures in a ‘reasonable, objective and impartial manner’, in sectors where specific commitments are undertaken.
So that measures relating to qualification requirements, technical standards and licensing ‘do not constitute unnecessary barriers to trade’, the CTS is empowered under GATS art VI:4 to develop any necessary disciplines, which must be (a) ‘based on objective and transparent criteria’, (b) ‘not more burdensome than necessary to ensure the quality of the service’ and (c) if licensing procedures are required, ‘not in themselves [be] a restriction on the supply of the service’.
While the CTS has so far not enacted binding disciplines under art VI:4, members cannot impose licensing and qualification requirements that impair their existing commitments and could not reasonably have been expected by other members when the commitments were made.
The CTS thus has the ability to influence the direction of blockchain, if it decides to set certain criteria relating to blockchain-based services and service suppliers. With the ability of this provision to impact on members’ ability to regulate services, the CTS is well-positioned to curb unreasonable and burdensome measures taken by members to prohibit or restrain access to blockchain platforms. Although this power has not been exercised before, this is a matter the CTS should seriously consider.
GATS permits members to adopt measures that would otherwise violate their commitments, provided that they pursue other policy goals and are not disguised trade restrictions. In an exhaustive list found in art XIV, GATS lists some ‘safety valves’ — domestic policy interests that allow members to deviate from their commitments. GATS art XIV is a combination of a chapeau and a list of specific public policy objectives. A two-tier analysis is involved: first, whether the measure falls under the listed exceptions, and secondly, whether the measure satisfies the chapeau.
Without going through the entire list, one possible justification that members may use to prohibit blockchain is art XIV(c): that such a measure is ‘necessary to secure compliance with laws or regulations which are not inconsistent with the provisions of [GATS]’. This provision indicates four types of problems that might be considered: deceptive and fraudulent practices, default on services contract, privacy of individuals and safety.
Article XIV(c) is broad in application and would likely cover the policy interests of countries that have banned blockchain, specifically cryptocurrencies. Most countries cite security or economic reasons for banning cryptocurrencies, as it would supposedly be easier for individuals to engage in money laundering or terrorist financing, or to transfer money outside their borders without financial institutions earning interest. Without an asset base, cryptocurrencies supposedly threaten global financial stability. For China, banning cryptocurrencies is a way to clean the financial sector, as cryptocurrencies supposedly encourage shadow banking, among other illicit activities. A central theme is the discomfort of central banks with a currency other than fiat money. Environmental reasons have also been cited as a policy objective, considering the vast amount of resources that cryptocurrency mining consumes. The protection of privacy, particularly the European Union’s General Data Protection Regulation’s (‘GDPR’) recognition of the ‘right to be forgotten’, may also potentially conflict with the decentralised and immutable features of blockchain.
In analysing whether members’ measures prohibiting blockchain run counter to art XVI(c), the analysis would involve the following steps determined by the Panel in US — Gambling: first, an assessment of whether the relevant measure prohibiting blockchain is ‘necessary’ to secure compliance with laws or regulations which are not in themselves inconsistent with GATS (ie the necessity test); and secondly, an assessment of whether the measure satisfies the chapeau.
Under the first step, the Panel in US — Gambling, adopting the factors considered in the GATT case US — Measures Relating to Shrimp from Thailand, assessed: (a) the ‘importance of the values or objectives the law or regulation is intended to protect’, (b) ‘the extent to which the measure contributes to the realization of the end pursued’, that is, to secure compliance with the law or regulation and (c) ‘the restrictive trade impact of the measure’. An important nuance is that ‘a measure can be said to be designed “to secure compliance” even if the measure cannot be guaranteed to achieve its result with absolute certainty’. In evaluating the restrictive trade impact, a claimant may also demonstrate other possible alternatives that will survive the necessity and chapeau tests, are less trade-restrictive and are reasonably available.
Under the chapeau test, the measure should not be ‘applied in a manner which would constitute a means of arbitrary or unjustifiable discrimination between countries where like conditions prevail, or a disguised restriction on trade in services’. Even where a measure satisfies the first limb of art XVI(c), the chapeau imposes additional limitations. The focus of the chapeau is on the ‘consistent application or enforcement of the measure’, preventing its assertion in an abusive manner. The member invoking the measure must show that it is not applied in a discriminatory or restrictive manner, based on the consistency of the measure’s application to foreign and domestic service suppliers, the pattern of enforcement and the number of service suppliers within the member’s jurisdiction.
Tackling the legitimacy of policy interests relating to blockchain in detail through the lens of GATS art XIV(c) is not the goal of this article. Doing so would require a thorough analysis of each measure. A faithful application of the tests under art XIV would involve an assessment of the importance of the many policy interests countries claim to have: economic development, security, privacy and environmental protection. Measures prohibiting Bitcoin or mining (both for Bitcoins and smart contracts) must contribute to the realisation of such policies (under the first limb) and must not be applied in a discriminatory manner or be a disguise restriction on trade (under the second limb). The material contribution of the measures to the achievement of the declared policies must be scrutinised, as the connection may turn out to be tenuous and grounded on a lack of understanding of how blockchain works.
One measure is the prohibition of cryptocurrencies because of their use for illicit operations, even when there is no direct evidence that Bitcoin has been steadily used for illicit operations. The 2017 UK National Risk Assessment on Money Laundering reported that in 2015, the risks associated with cryptocurrencies ‘were assessed to be low for both money laundering and terrorist financing’, and that there remains little evidence that they are being used for either. Meanwhile, the same crimes continue to proliferate in banks and other mainstream financial institutions, without the same wholesale prohibitions imposed on cryptocurrencies. The GATS Annex on Financial Services further complicates the analysis. It provides, inter alia, that a ‘Member shall not be prevented from taking measures for prudential reasons ... to ensure the integrity and stability of the financial system’. Under the guise of prudential measures, members may in principle be allowed to prohibit the use of cryptocurrencies (even if the same measures are not applied to other financial services) as an exception to its commitments. However, adopting such measures is hardly straightforward, practical or reasonable on an international level. Considering the interrelatedness and internationalisation of financial services, including blockchain-powered cryptocurrencies, members must also recognise that an integrated international approach is needed for economic advancement and financial stability.
Another claim is that the immutability and transparency of the blockchain ledger poses privacy risks and that, therefore, measures prohibiting blockchain are consistent with privacy regulations. To be clear, privacy and cybersecurity are important policy considerations in safeguarding the internet. This claim, however, is not ironclad, as many blockchain platforms, including Bitcoin and Ethereum, in fact anonymise their users’ identities by using hashing and asymmetric cryptography. In order to be justified under GATS art XIV(c), a member must prove that such a prohibitive measure is necessary to comply with privacy laws, and contributes to the realisation of privacy goals. The GDPR ‘regulates the processing by an individual ... or an organisation of personal data relating to individuals in the EU’. If the blockchain platform does not process personally identifiable information in the first place, such a measure can hardly be justified as securing compliance with the GDPR. On the other hand, if there is a way for participants in the blockchain platform to piece together certain datasets and de-anonymise such datasets to trace them to a particular user, and there are no adequate protections in place, such a measure may be warranted after all. The chapeau test is another matter altogether under the two-tier test: the member invoking the measure must show that it is not being applied in a discriminatory or restrictive manner. The use of GATS art XIV to justify the prohibition of blockchain on the ground of privacy-related concerns is a complex undertaking that will require WTO tribunals to consider, among other factors, the technical feasibility of such prohibition, its operation and the possible alternatives that might achieve a similar level of privacy protection.
Moreover, several alternatives that are less trade restrictive than a wholesale prohibition of blockchain or cryptocurrencies are arguably reasonably available and may achieve the same policy objectives. Some countries such as the Philippines have issued licensing regulations for cryptocurrency exchanges. The European Union Committee on International Trade has called for a motion, among others, recognising that blockchain and the GDPR are ‘underpinned by common principles of ensuring secured and self-governed data’ — contrary to its supposed privacy violation.
Although this article has argued for the adaptability of GATS in liberalising blockchain, the fact remains that several members made their commitments based on a pre-internet classification of services. Based on an understanding of how blockchain works, its classification as a data processing service under computer and related services under the CPC (1991) certainly makes sense. However, members that prohibit blockchain will nonetheless argue against the application of their commitment under this service, or may excuse themselves based on the chapeau. Unless the WTO is faced with a case involving one member prohibiting blockchain on one hand and a member relying on blockchain for its digital trade transactions on the other, the application of GATS will remain open to question.
This part discusses existing parallel developments or possibilities that may either supplement GATS or provide another reliable legal framework for blockchain. Unfortunately, some of these developments may also fragment the consensus on the liberalisation of blockchain.
First, members may expressly agree to a set of principles expressly adopting GATS for digital trade, including far-reaching additional commitments in their respective schedules. Such an undertaking would obviously require political will across parties, but may be feasible for classifications with a relatively high level of commitment, such as computer and related services. Unfortunately, such an agreement clarifying the relevance of GATS to digital trade, including blockchain, is unlikely to materialise in the near future, because of the complexities of digital trade issues and the lack of political consensus regarding the correct regulatory approach.
A second path that can make GATS a better fit would be to push the digital trade agenda in the Trade in Services Agreement (‘TiSA’). With the support of economies comprising 70% of the international services trade, TiSA intends to liberalise services sectors through deeper market access commitments that can liberalise blockchain. One example is the Swiss proposal: with its proposal on Information and Communication Technology Services, Switzerland submitted detailed provisions on open networks, network access, transparent and independent regulatory agencies and cross-border information flows — all of which are important in developing blockchain technologies.
Lastly, the lack of a definitive path towards the liberalisation of trade in services under GATS has led many countries to seek more strategic solutions that advance their interests, specifically through bilateral and regional preferential trade agreements (‘PTAs’) with entire sections dedicated to digital trade issues and the cross-border delivery of services. Such PTAs often expand on GATS commitments and liberalisation. The US, following the collapse of the Doha Round in 2008, has used PTAs as a means of advancing its trade agenda in e-commerce. Bilateral PTAs between the US and Australia, Chile, Morocco and Singapore have incorporated e-commerce chapters, followed by a spread of bilateral PTAs not involving the US, such as India–Singapore, Japan–Singapore, Korea–Singapore, New Zealand–Singapore, Singapore–Australia and Thailand–New Zealand.
PTAs build on the GATS principles of liberalisation and democratisation, and additionally provide solutions to many of the grey areas in GATS, including a formalised definition of digital products, the explicit recognition that the WTO rules apply to e-commerce, the applicability of trade rules to the electronic supply of services as well as national treatment and MFN obligations for digital products. They also expressly recognise the technology neutrality principle elaborated in WTO case law and innovate, primarily by using a negative-list approach to broaden commitments to existing and future digital services and by removing local presence requirements.
PTAs are better at reconciling widely diverging interests, particularly in the classification of digital services and the reconciliation of localisation measures. They may also serve as a stepping stone for hashing out divisive issues and testing the effectiveness of digital trade solutions; then, once a critical mass is achieved, help build consensus on a multilateral level such as the WTO. As more Members agree, for example, on a definition of digital products in their respective PTAs, amending GATS to reflect this consensus is a step closer to reality. The deep level of commitments in such PTAs may then serve as a useful model in amending GATS as WTO negotiations push through. An example is the Trans-Pacific Partnership Agreement (‘TPP’), considered as ‘the most ambitious trade policy ever designed for the Internet and electronic commerce’. Concluded between Brunei Darussalam, Chile, New Zealand, Singapore and the US, the seven-year negotiation provided for deep commitments to digital trade issues, including a single, global internet that ensures cross-border data flows.
At the same time, though, PTAs may increasingly fragment the international approach to digital trade in services, given the heterogeneity of e-commerce provisions found in various PTAs. Further, the progress in digital trade achieved by PTAs may easily be retracted by the immediate withdrawal of its members. The TPP again provides a sobering example, with the withdrawal of US President Donald Trump in January 2017. Moreover, increasing reliance on PTAs to resolve digital trade issues may undermine the WTO dispute settlement mechanism and even international law, especially where the jurisdictions of WTO and PTA tribunals overlap. There has also been little use of the dispute settlement mechanism found in PTAs so far, indicating little follow-up activity after their conclusion.
This article has argued that GATS is flexible enough to accommodate blockchain. Whilst most of the recent movements on digital trade in services were done on the level of PTAs, and whilst replicating these ‘deep’ digital trade rules in GATS is a long-term exercise that will not likely happen in the foreseeable future, blockchain does not exist in a legal vacuum. Its reliance on cross-border data flows and infrastructure allows people from all over the world to supply their services in the global market, without the intervention of an all-too-powerful intermediary. At its core, blockchain is an idealisation of the democratised economic order aspired to by the WTO through GATS.
Applying GATS to blockchain means that many prohibitive or prudential regulations imposed by members — from banning mining to criminalising the possession of cryptocurrencies — may turn out to be trade barriers. Such measures must thus be weighed against existing commitments to data processing services (under computer and related services) that members have undertaken in their schedules. The members must, then, respect their GATS obligations: transparency, domestic regulation and MFN obligations apply without reservation, while national treatment and market access commitments must be examined based on what is set forth in their schedules. To the extent that such blockchain measures purport to pursue other policy goals, they may be justified under GATS art XIV.
Although new rules would admittedly be preferable to adapt to advancements in digital trade, and whilst GATS has arguably not kept pace with technological developments in an unambiguous manner, GATS does fill the need for a supportive policy. Obviously more needs to be done by the WTO, such as a multi-stakeholder and multi-layered engagement focusing on digital trade, preferably including the blockchain agenda. Countries that have an interest in the growth of blockchain and intend to be at the forefront of this technology must take an active role in setting international standards for its development.
 Joshua P Meltzer and Peter Lovelock, ‘Regulating for a Digital Economy: Understanding the Importance of Cross-Border Data Flows in Asia’ (Working Paper No 113, Global Economy & Development, Brookings Institution, March 2018) iii.
 Julie Maupin, ‘The G20 Countries Should Engage with Blockchain Technologies to Build an Inclusive, Transparent, and Accountable Digital Economy for All’ (Policy Brief, G20 Insights, 16 March 2016) 2–3.
 Emma McClarkin, European Parliament Committee on International Trade, Draft Report on Blockchain: A Forward-Looking Trade Policy (Draft Report No 2018/2085(INI), 18 July 2018) 8.
 Mike Orcutt, ‘Q: How Secure Is Blockchain Really? A: It Turns out “Secure” Is a Funny Word to Pin Down’ (2018) 121(3) MIT Technology Review 40.
 Lee Tuthill and Martin Roy, ‘GATS Classification Issues for Information and Communication Technology Services’ in Mira Burri and Thomas Cottier (eds), Trade Governance in the Digital Age: World Trade Forum (Cambridge University Press, 2012) 157, 158.
 Marrakesh Agreement Establishing the World Trade Organization, opened for signature 15 April 1994, 1867 UNTS 3 (entered into force 1 January 1995) annex 1B (‘General Agreement on Trade in Services’) (‘GATS’).
 Satoshi Nakamoto, Bitcoin: A Peer-to-Peer Electronic Cash System (White Paper, 31 October 2008) 1 <https://bitcoin.org/bitcoin.pdf>, archived at <https://perma.cc/5W87-YLYG>.
 Maupin (n 2) 2.
 William Mougayar, The Business Blockchain: Promise, Practice, and Application of the Next Internet Technology (Wiley, 2016) 21.
 Nakamoto (n 7) 6.
 Hashing refers to the conversion of a transaction input of any given length into a fixed output length: ibid 3.
 Mougayar (n 10) 10.
 Harish Natarajan, Solvej Krause and Helen Gradstein, ‘Distributed Ledger Technology (DLT) and Blockchain’ (FinTech Note No 1, World Bank Group, 1 January 2017) 14 <http://documents.worldbank.org/curated/en/134831513333483951/pdf/WP-PUBLIC-Distributed-Ledger-Technology-and-Blockchain-Fintech-Notes.pdf> , archived at <https://perma.cc/GK2M-LQBV>.
 UK Government Chief Scientific Adviser, Distributed Ledger Technology: Beyond Block Chain (Report, 19 January 2016) 18 <https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/492972/gs-16-1-distributed-ledger-technology.pdf>, archived at <https://perma.cc/LGY8-KFKG> (‘Distributed Ledger Technology Report’).
 Jeff Desjardins, ‘The Power of Smart Contracts on the Blockchain’, Visual Capitalist (online, 24 October 2017) <http://www.visualcapitalist.com/smart-contracts-blockchain/> , archived at <https://perma.cc/ZHP8-VLG2>.
 Saifedean Ammous, ‘Economics beyond Financial Intermediation: Digital Currencies’ Possibilities for Growth, Poverty Alleviation, and International Development’ (2015) 30(3) Journal of Private Enterprise 19, 35.
 Desjardins (n 16); Natarajan, Krause and Gradstein (n 14) 15.
 Desjardins (n 16).
 Tien Tuan Anh Dinh et al, ‘Untangling Blockchain: A Data Processing View of Blockchain Systems’ (2018) 30(7) IEEE Transactions on Knowledge and Data Engineering 1366, 1369.
 Natarajan, Krause and Gradstein (n 14) 1.
 Malcolm Campbell-Verduyn and Marcel Goguen, ‘The Mutual Constitution of Technology and Global Governance: Bitcoin, Blockchains, and the International Anti-Money-Laundering Regime’ in Malcolm Campbell-Verduyn (ed), Bitcoin and Beyond: Cryptocurrencies, Blockchains, and Global Governance (Routledge, 2018) 69, 74; Minhaj Ahmad Khan and Khaled Salah, ‘IoT Security: Review, Blockchain Solutions, and Open Challenges’ (2018) 82 Future Generation Computer Systems 395.
 Maupin (n 2).
 Natarajan, Krause and Gradstein (n 14) 11.
 Kai Jia and Falin Zhang, ‘Between Liberalization and Prohibition: Prudent Enthusiasm and the Governance of Bitcoin/Blockchain Technology’ in Malcolm Campbell-Verduyn (ed), Bitcoin and Beyond: Cryptocurrencies, Blockchains, and Global Governance (Routledge, 2018) 88.
 Ibid 104.
 Zhuling Chen, ‘How Should We Regulate Blockchain? It Depends on Which Country You Ask’, Fortune (online, 25 June 2018) <http://fortune.com/2018/06/25/blockchain-cryptocurrency-technology-regulation-bitcoin-ethereum/> , archived at <https://perma.cc/676H-7M3S>.
 Stan Higgins, ‘German Central Bank Tests Blockchain Trading Prototype’, Coindesk (online, 28 November 2016) <https://www.coindesk.com/german-central-bank-blockchain-trading/>, archived at <https://perma.cc/QAE4-ARZW>.
 Chen (n 27).
 Distributed Ledger Technology Report (n 15).
 See Mauro Sacramento, ‘Authorities: Ukraine Has No Plans to Regulate Cryptocurrency Mining’, CCN (online, 27 June 2018) <https://www.ccn.com/authorities-ukraine-has-no-plans-to-regulate-cryptocurrency-mining/>, archived at <https://perma.cc/53LK-GFHJ>; Nerses Isajanyan, ‘Regulation of Cryptocurrency: Belarus’, Library of Congress (Web Page, June 2018) <https://www.loc.gov/law/help/cryptocurrency/belarus.php#_ftn2>, archived at <https://perma.cc/78BD-6CZJ>.
 Becky Leighton, ‘Borders Blocking Bitcoin: Countries Who Are Still Anti-Cryptocurrency’, Coin Insider (online, 27 May 2018) <https://www.coininsider.com/countries-banning-bitcoin/>, archived at <https://perma.cc/P5Y3-6AFL>.
 Allen Scott, ’11 Countries Where Bitcoin Is Still Illegal’, Bitcoinist (online, 18 April 2018) <https://bitcoinist.com/11-countries-bitcoin-still-illegal/>, archived at <https://perma.cc/PL3G-5T5D>; ‘Cryptocurrencies by Countries’, Thomson Reuters (Blog Post, 25 October 2017) <https://blogs.thomsonreuters.com/answerson/world-cryptocurrencies-country/>, archived at <https://perma.cc/CP9F-JL3E>.
 Leighton (n 32).
 Jerin Mathew, ‘Ecuador to Create Government-Run Digital Currency as It Bans Bitcoin’, International Business Times (online, 25 July 2014) <https://www.ibtimes.co.uk/ecuador-create-government-run-digital-currency-it-bans-bitcoin-1458280>, archived at <https://perma.cc/8NF6-R54X>.
 Leighton (n 32).
 ‘Gibraltar Is the First Regulated Cryptocurrency Country’, Oracle Times (online, 20 April 2018) <https://oracletimes.com/gibraltar-is-the-first-regulated-cryptocurrency-country/>, archived at <https://perma.cc/2ALF-EGVX>.
 ‘Virtuaalivaluuttojen Verotus’ [Income Tax on Virtual Currency], Vero Skatt [Finland Tax Authority] (Web Page, 29 May 2018) <https://www.vero.fi/en/detailed-guidance/guidance/48411/taxation-of-virtual-currencies/>, archived at <https://perma.cc/5DZC-JUVE>.
 Nestor A Espenilla Jr, Central Bank of the Philippines, Guidelines for Virtual Currency (VC) Exchanges 2017 (Circular No 944 of Series of 2017, 6 February 2017) 1–6 (‘Guidelines for Virtual Currency Exchanges’).
 See Daniel Ren, ‘Why the Bitcoin Has Become China’s New Darling’, South China Morning Post (online, 8 June 2016) <https://www.scmp.com/news/china/economy/article/1968733/why-bitcoin-has-become-chinas-new-darling>, archived at <https://perma.cc/W93M-CRJ3>.
 See Zheping Huang, ‘China, Home to the World’s Biggest Cryptocurrency Mining Farms, Now Wants to Ban Them Completely’, South China Morning Post (online, 9 April 2019 <https://www.scmp.com/tech/policy/article/3005334/china-home-worlds-biggest-cryptocurrency-mining-farms-now-wants-ban>, archived at <https://perma.cc/DV4D-8K4S>.
 See Evelyn Cheng, ‘Chinese President Xi Jinping Calls Blockchain a “Breakthrough” Technology’, CNBC (online, 30 May 2018) <https://www.cnbc.com/2018/05/30/chinese-president-xi-jinping-calls-blockchain-a-breakthrough-technology.html>, archived at <https://perma.cc/7UCE-UTHZ>.
 Chen Jia and Ren Xiaojin, ‘PBOC Gets Tougher on Bitcoin’, China Daily (online, 5 January 2018) <http://usa.chinadaily.com.cn/a/201801/05/WS5a4eb4cba31008cf16da527c.html> , archived at <https://perma.cc/3N9L-PAP2>.
 Brenda Goh and Alun John, ‘China Wants to Ban Bitcoin Mining’, Thomson Reuters (online, 9 April 2019) <https://www.reuters.com/article/us-china-cryptocurrency/china-wants-to-ban-bitcoin-mining-idUSKCN1RL0C4>, archived at <https://perma.cc/F3BL-GAWU>.
 Massimo Di Giuda, ‘Countries Where the Cryptocurrencies Are Banned: Busted for Bitcoin’, Bitnews Today (online, 25 April 2018) <https://bitnewstoday.com/market/bitcoin/countries-where-the-cryptocurrencies-are-banned-busted-for-bitcoin/>, archived at <https://perma.cc/DV7W-N8LM>.
 Jia and Xiaojin (n 43).
 Bloomberg, ‘This Is How China Is Stifling Bitcoin and Cryptocurrencies’, Fortune (online, 17 January 2018) <http://fortune.com/2018/01/17/china-bitcoin-cryptocurrency-crackdown/> , archived at <https://perma.cc/3X7V-LHN7>.
 See Maupin (n 2).
 Ibid 2.
 Ammous (n 17) 43.
 Nick Spanos, ‘Stifling Innovation with Regulation: Why Countries Shouldn’t Ban Cryptocurrency Trading’, The Block (online, 1 February 2018) <https://www.blockchaintechnology-news.com/2018/02/01/stifling-innovation-regulation-countries-shouldnt-ban-cryptocurrency-trading/>, archived at <https://perma.cc/93NG-6EBJ>.
 See generally Daniela Sonderegger, ‘A Regulatory and Economic Perplexity: Bitcoin Needs Just a Bit of Regulation’ (2015) 47 Washington University Journal of Law and Policy 175. See also Loi Luu, ‘Blockchain Adoption: How Close Are We Really?’, Forbes (online, 26 January 2018) <https://www.forbes.com/sites/luuloi/2018/01/26/blockchain-adoption-how-close-are-we-really/>, archived at <https://perma.cc/6YC6-79LZ>.
 Marrakesh Agreement Establishing the World Trade Organization, opened for signature 15 April 1994, 1867 UNTS 3 (entered into force 1 January 1995) annex 1A (‘General Agreement on Tarriffs and Trade’) (‘GATT’).
 Michael Hart and Ramesh Chaitoo, ‘Electronic Commerce and International Trade Rules’ (1999) 2(6) Journal of World Intellectual Property 911, 915.
 Ibid 917–18.
 See Simon Lester, Bryan Mercurio and Arwel Davies, World Trade Law: Text, Materials and Commentary (Hart Publishing, 2nd ed, 2012) 636.
 See, eg, Appellate Body Report, European Communities — Measures Affecting Asbestos and Asbestos-Containing Products, WTO Doc WT/DS135/AB/R (12 March 2001) , .
 Mira Burri, ‘The International Economic Law Framework for Digital Trade’ (2015) 134(2) Zeitschrift für Schweizerisches Recht 7, 18 (‘Digital Trade Framework’).
 Sacha Wunsch-Vincent, ‘Trade Rules for the Digital Age’ in Marion Panizzon, Nicole Pohl and Pierre Sauvé (eds), GATS and the Regulation of International Trade in Services (Cambridge University Press, 2008) 497, 502–3 (‘Trade Rules for the Digital Age’).
 Burri, ‘Digital Trade Framework’ (n 58) 18.
 Hart and Chaitoo (n 54) 915.
 Burri, ‘Digital Trade Framework’ (n 58) 38–9.
 Work Programme on Electronic Commerce, WTO Doc WT/L/274 (30 September 1998, adopted 25 September 1998) paras 1.3, 2.1.
 Panel Report, United States — Measures Affecting the Cross-Border Supply of Gambling and Betting Services, WTO Doc WT/DS285/R (10 November 2004) (‘US — Gambling’).
 Sacha Wunsch-Vincent, ‘The Internet, Cross-Border Trade in Services, and the GATS: Lessons from US — Gambling’ (2006) 5(3) World Trade Review 319, 323 (‘Lessons from US — Gambling’).
 Panel Report, US — Gambling, WTO Doc WT/DS285/R (n 65) [6.285].
 Appellate Body Report, China — Measures Affecting Trading Rights and Distribution Services for Certain Publications and Audiovisual Entertainment Products, WTO Doc WT/DS363/AB/R (21 December 2009) (‘China — Publications and Audiovisual Products’).
 Ibid –. See also Henry Gao, ‘Google’s China Problem: A Case Study on Trade, Technology and Human Rights under the GATS’ (2011) 6(2) Asian Journal of WTO and International Health Law and Policy 349, 368–9 (‘Google’s China Problem’).
 Burri, ‘Digital Trade Framework’ (n 58) 39–41.
 Services Sectoral Classification List, GATT Doc MTN.GNS/W/120 (10 July 1991) (Note by the Secretariat) (‘GATS Classification List’).
 Statistical Office of the United Nations, Department of International Economic and Social Affairs, Provisional Central Product Classification, UN Doc ST/ESA/STAT/SER.M/77 (1991) (‘CPC (1991)’). See below Part IV(A).
 Burri, ‘Digital Trade Framework’ (n 58) 34.
 Gao, ‘Google’s China Problem’ (n 69) 351.
 Burri, ‘Digital Trade Framework’ (n 58) 34.
 Shin-yi Peng, ‘GATS and the Over-the-Top Services: A Legal Outlook’ (2016) 50(1) Journal of World Trade 21, 22–3.
 Gao, ‘Google’s China Problem’ (n 69) 366.
 Statistics Division, Department of Economic and Social Affairs, United Nations, Central Product Classification (CPC): Version 2.1, UN Doc ST/ESA/STAT/SER.M/77/Ver.2.1 (2015) (‘CPC (2015)’).
 See Peng (n 76) 27.
 Appellate Body Report, United States — Measures Affecting the Cross-Border Supply of Gambling and Betting Services, WTO Doc WT/DS285/AB/R (7 April 2005)  (‘US — Gambling’).
 Peng (n 76) 28.
 Nellie Munin, Legal Guide to GATS (Kluwer Law International, 2010) 52.
 Ibid 55.
 Ibid 52.
 Nakamoto (n 7) 4.
 GATS (n 6) art XXVIII(b).
 See Distributed Ledger Technology Report (n 15) 18.
 See IBM Research, ‘International Trade Solution on Blockchain’ (YouTube, 16 May 2016) 00:00:56–00:01:54 <https://www.youtube.com/watch?v=r0LsnzAe1Yg>.
 Ibid 00:06:00–00:06:24.
 GATS (n 6) art XXVIIII(i).
 Ibid art XXVIIII(ii).
 Munin (n 82) 74.
 GATS (n 6) art XXVIII(a).
 World Trade Organization, ‘WTO Disputes Reach 400 Mark’ (Press Release Press/578, 6 November 2009) <https://www.wto.org/english/news_e/pres09_e/pr578_e.htm>, archived at <https://perma.cc/67HM-8SUR>.
 GATS (n 6) art I:2.
 Working Party of the Trade Committee, Organisation for Economic Co-operation and Development, Regulation of Services Traded Electronically (Final Report No TD/TC/WP(2002)32/FINAL, 11 December 2002) 9–10 .
 While electronic supply can also be done under Modes 3 and 4, barriers to trade and jurisdictional issues under Modes 1 and 2 are less clearly delineated, because Modes 3 and 4 presume a physical presence, through a legal entity or natural persons, in the territory of the consumer: ibid 6 –.
 See Wunsch-Vincent, ‘Lessons from US — Gambling’ (n 66) 324.
 Joshua P Meltzer, ‘A New Digital Trade Agenda’ (Overview Paper, E15 Initiative, August 2015) 10.
 Ibid; Sacha Wunsch-Vincent, WTO, E-Commerce, and Information Technologies: From the Uruguay Round through the Doha Development Agenda, ed Joanna McIntosh (UN ICT Task Force, 2004) 86 .
 Panel Report, US — Gambling, WTO Doc WT/DS285/R (n 65) [6.29]–[6.32].
 Appellate Body Report, US — Gambling, WTO Doc WT/DS285/AB/R (n 80) .
 Wunsch-Vincent, ‘Lessons from US — Gambling’ (n 66) 326.
 Chris Jaikaran, ‘Blockchain: Background and Policy Issues’ (CRS Report No R45116, Congressional Research Service, 28 February 2018) 4–5 <https://fas.org/sgp/crs/misc/R45116.pdf>, archived at <https://perma.cc/5WF8-VUUF>.
 Ibid 4.
 Jacob Funk Kirkegaard, ‘Distance Isn’t Quite Dead: Recent Trade Patterns and Modes of Supply in Computer and Information Services in the United States and NAFTA Partners’ (Working Paper No 08–10, Peterson Institute for International Economics, October 2008) 4.
 Munin (n 82) 140–1.
 Panel Report, China — Measures Affecting Trading Rights and Distribution Services for Certain Publications and Audiovisual Entertainment Products, WTO Doc WT/DS363/R (12 August 2009) [7.1165].
 Ibid [7.1218]; Appellate Body Report, China — Publications and Audiovisual Products, WTO Doc WT/DS363/AB/R (n 68) , .
 Appellate Body Report, China — Publications and Audiovisual Products, WTO Doc WT/DS363/AB/R (n 68) 
 Peng (n 76) 29.
 Appellate Body Report, United States — Import Prohibition of Certain Shrimp and Shrimp Products, WTO Doc WT/DS58/AB/R (12 October 1998)  (‘US — Shrimp’), cited in China — Publications and Audiovisual Products, WTO Doc WT/DS363/AB/R (n 68)  n 705.
 Appellate Body Report, US — Shrimp, WTO Doc WT/DS58/AB/R (n 113) .
 Appellate Body Report, China — Publications and Audiovisual Products, WTO Doc WT/DS363/AB/R (n 68) –.
 See Tuthill and Roy (n 5) 177.
 See ibid.
 Amy Porges and Alice Enders, ‘Data Moving across Borders: The Future of Digital Trade Policy’ (Think Piece, E15 Initiative, April 2016) 9.
 Ines Willemyns, ‘The GATS (In)Consistency of Barriers to Digital Services Trade’ (Working Paper No 207, Leuven Centre for Global Governance Studies, September 2018) 5.
 Tuthill and Roy (n 5) 159.
 Willemyns (n 119) 5.
 Mireille Cossy, ‘Some Thoughts on the Concept of “Likeness” in the GATS’ in Marion Panizzon, Nicole Pohl and Pierre Sauvé (eds), GATS and the Regulation of International Trade in Services (Cambridge University Press, 2008) 327, 334.
 Tuthill and Roy (n 5) 177.
 See Henry Gao, ‘Googling for the Trade–Human Rights Nexus in China: Can the WTO Help?’ in Mira Burri and Thomas Cottier (eds), Trade Governance in the Digital Age: World Trade Forum (Cambridge University Press, 2012) 247, 251–3 (‘Googling for the Trade’).
 Cossy (n 122) 334.
 Peng (n 76) 27–8.
 Wunsch-Vincent, ‘Trade Rules for the Digital Age’ (n 59) 502.
 Andrew D Mitchell and Neha Mishra, ‘Data at the Docks: Modernizing International Trade Law for the Digital Economy’ (2018) 20(4) Vanderbilt Journal of Entertainment and Technology Law 1073, 1126–7.
 CPC (1991), UN Doc ST/ESA/STAT/SER.M/77 (n 72) 54–5.
 Gao, ‘Googling for the Trade’ (n 124) 254.
 CPC (1991), UN Doc ST/ESA/STAT/SER.M/77 (n 72) 64.
 Damien Cosset, ‘Blockchain: What Is in a Block?’, DEV (Forum Post, 27 December 2017) <https://dev.to/damcosset/blockchain-what-is-in-a-block-48jo>, archived at <https://perma.cc/BZ44-H824>.
 Maryanne Murray, ‘Blockchain Explained’, Reuters Graphics (Web Page, 15 June 2018) <http://graphics.reuters.com/TECHNOLOGY-BLOCKCHAIN/010070MF1E7/index.html> , archived at <https://perma.cc/3E4A-5UXY>.
 Edward Burton, ‘5 Ways in Which Blockchain Is Not Just a Slow Database’, Medium (online, 24 May 2018) <https://medium.com/@chainfrog/5-reasons-that-blockchain-is-not-just-a-slow-database-55fe9d913578>, archived at <https://perma.cc/DU9G-2T82>; Nolan Bauerle, ‘What Is the Difference between a Blockchain and a Database?’, CoinDesk (online) <https://www.coindesk.com/information/what-is-the-difference-blockchain-and-database/>, archived at <https://perma.cc/7LEN-C3XW>.
 Bauerle (n 134).
 Burton (n 134).
 Computer and Related Services, WTO Doc S/C/W/45 (14 July 1998) (Background Note by the Secretariat) .
 Telecommunications Services, WTO Doc S/C/W/299 (10 June 2009) (Background Note by the Secretariat) .
 GATS (n 6) annex (‘Annex on Telecommunications’) art 3(a).
 Work Program on Electronic Commerce: Ensuring that Trade Rules Support Innovative Advances in Computer Applications and Platforms, Such as Mobile Applications and the Provision of Cloud Computing Services, WTO Doc S/C/W/339 (20 September 2011) (Communication from the United States on 19 September 2011) .
 CPC (1991), UN Doc ST/ESA/STAT/SER.M/77 (n 72) 64.
 Gao, ‘Googling for the Trade’ (n 124) 255.
 Nakamoto (n 7) 3.
 Cosset (n 132).
 Jaikaran (n 105) 4.
 Ibid; Nakamoto (n 7) 3–4.
 Nakamoto (n 7) 3.
 Rui Roriz and José Luís Pereira, ‘IoT Applications Using Blockchain and Smart Contracts’ in Tatiana Antipova and Alvaro Rocha (eds), Digital Science (Springer, 2019) 426, 428.
 Jaikaran (n 105) 4.
 Ibid 4–5.
 Natarajan, Krause and Gradstein (n 14) 2.
 Jaikaran (n 105) 7.
 Panel Report, China — Certain Measures Affecting Electronic Payment Services, WTO Doc WT/DS413/R (16 July 2012) [7.100], [7.111].
 Ibid [7.99].
 CPC (2015), UN Doc ST/ESA/STAT/SER.M/77/Ver.2.1 (n 78). See also Peng (n 76) 27.
 Mitchell and Mishra (n 128) 1126–7.
 Mira Burri, ‘The Governance of Data and Data Flows in Trade Agreements: The Pitfalls of Legal Adaptation’ (2017) 51(1) UC Davis Law Review 65, 72, 80 (‘The Governance of Data’).
 Ibid 72.
 Munin (n 82) 183.
 Ibid 105.
 Ibid 122, citing Panel Report, EC Measures concerning Meat and Meat Products (Hormones) — Complaint by the United States, WT/DS26/R/USA (18 August 1997) [4.261]–[4.266].
 Cossy (n 122) 354.
 See Jaikaran (n 105) 5–8.
 GATS (n 6) art XVII.
 Ibid art XVII:1.
 Munin (n 82) 160.
 GATS (n 6) art XVII:2.
 Panel Report, Canada — Certain Measures Affecting the Automotive Industry, WTO Docs WT/DS139/R and WT/DS142/R (11 February 2000) [10.84].
 Venzen Khaosan, ‘Ecuador Bans Bitcoin in Favor of Own National Cryptocurrency’, CCN (online, 27 July 2014) <https://www.ccn.com/ecuador-bans-bitcoin-favor-own-national-cryptocurrency/>, archived at <https://perma.cc/PTU2-JJW9>.
 Munin (n 82) 27.
 Ibid 183.
 Lester, Mercurio and Davies (n 56) 654.
 Panel Report, US — Gambling, WTO Doc WT/DS285/R (n 65) [6.361].
 Appellate Body Report, US — Gambling, WTO Doc WT/DS285/AB/R (n 80) –.
 Panel Report, US — Gambling, WTO Doc WT/DS285/R (n 65) [6.362]–[6.365].
 For an overview of countries where Bitcoin is prohibited and criminalised, see Leighton (n 32).
 GATS (n 6) art III.
 Ibid art III:4.
 Ibid art VI:1.
 Ibid art VI:4.
 Ibid art VI:5.
 Lester, Mercurio and Davies (n 56) 642.
 Burri, ‘The Governance of Data’ (n 159) 88.
 Markus Krajewski, National Regulation and Trade Liberalization in Services: The Legal Impact of the General Agreement on Trade in Services (GATS) on National Regulatory Autonomy (Kluwer Law International, 2003) 196.
 Munin (n 82) 337. See also Thomas Cottier, Panos Delimatsis and Nicolas Dieblod, ‘Article XIV GATS: General Exceptions’ in Rüdiger Wolfrumm, Peter-Tobias Stoll and Clemens Feinäugle (eds), WTO — Trade in Services (Martinus Nijhoff, 2008) vol 6, 291.
 Munin (n 82) 340.
 Lester, Mercurio and Davies (n 56) 367.
 GATS (n 6) art XIV(c).
 Lester, Mercurio and Davies (n 56) 371, 385.
 Di Giuda (n 45).
 Joe Sommerlad, ‘Cryptocurrencies Should Be Banned Because Regulating Them Is Too Difficult, Says Indian Economic Expert’, The Independent (online, 12 March 2018) <https://www.independent.co.uk/life-style/gadgets-and-tech/news/bitcoin-price-latest-india-cryptocurreny-ban-proposal-regulation-registration-blockchains-a8251726.html>, archived at <https://perma.cc/96QS-7J36>.
 Bloomberg (n 47).
 Darrell M West and Jack Karsten, ‘Cryptocurrency Crackdowns Deny a Potential Source of Innovation’, Brookings (online, 30 October 2017) <https://www.brookings.edu/blog/techtank/2017/10/30/cryptocurrency-crackdowns-deny-a-potential-source-of-innovation/>, archived at <https://perma.cc/P3TM-323U>.
 Jon Russell, ‘China Is Reportedly Moving to Clamp Down on Bitcoin Miners’, TechCrunch (Blog Post) <http://social.techcrunch.com/2018/01/08/china-is-reportedly-moving-to-clampdown-on-bitcoin-miners/> , archived at <https://perma.cc/Z4NM-AK9H>.
 Chen (n 27).
 Panel Report, US — Gambling, WTO Doc WT/DS285/R (n 65) [6.446], [6.449].
 Appellate Body Report, United States — Measures Relating to Shrimp from Thailand; United States — Customs Bond Directive for Merchandise Subject to Anti-Dumping/Countervailing Duties, WTO Docs WT/DS343/AB/R and WT/DS345/AB/R (16 July 2008) .
 Ibid; Panel Report, US — Gambling, WTO Doc WT/DS285/R (n 65) [6.557].
 Appellate Body Report, Mexico — Tax Measures on Soft Drinks and Other Beverages, WTO Doc WT/DS308/AB/R (6 March 2006) .
 Appellate Body Report, US — Gambling, WTO Doc WT/DS285/AB/R (n 80) .
 Appellate Body Report, Korea — Measures Affecting Imports of Fresh, Chilled and Frozen Beef, WTO Docs WT/DS161/AB/R and WT/DS169/AB/R (11 December 2000) , –.
 GATS (n 6) art XIV.
 Lester, Mercurio and Davies (n 56) 371, 373.
 Cottier, Delimatsis and Dieblod (n 187) 321.
 Munin (n 82) 373.
 See Campbell-Verduyn and Goguen (n 22) 75.
 HM Treasury and Home Office, National Risk Assessment of Money Laundering and Terrorist Financing 2017 (Policy Paper, 26 October 2017) 38 <https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/655198/National_risk_assessment_of_money_laundering_and_terrorist_financing_2017_pdf_web.pdf>, archived at <https://perma.cc/98RZ-TLTT>.
 HM Treasury and Home Office, UK National Risk Assessment of Money Laundering and Terrorist Financing (Policy Paper, 15 October 2015) 10, 12 <https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/468210/UK_NRA_October_2015_final_web.pdf>, archived at <https://perma.cc/6RQH-DYRB>.
 GATS (n 6) annex (‘Annex on Financial Services’) art 2(a).
 Munin (n 82) 416.
 Andrew D Mitchell, Jennifer K Hawkins and Neha Mishra, ‘Dear Prudence: Allowances under International Trade and Investment Law for Prudential Regulation in the Financial Services Sector’ (2016) 19(4) Journal of International Economic Law 787, 814.
 Mitchell and Mishra (n 128) 1094.
 Mougayar (n 10) 49.
 ‘What Does the General Data Protection Regulation (GDPR) Govern?’, European Commission: Policies, Information and Services (Web Page) <https://ec.europa.eu/info/law/law-topic/data-protection/reform/what-does-general-data-protection-regulation-gdpr-govern_en>, archived at <https://perma.cc/USV3-5AQX>.
 See Appellate Body Report, United States — Standards for Reformulated and Conventional Gasoline, WTO Doc WT/DS2/AB/R (29 April 1996) 23.
 Mitchell and Mishra (n 128) 1094, citing Andrew D Mitchell and Jarrod Hepburn, ‘Don’t Fence Me In: Reforming Trade and Investment Law to Better Facilitate Cross-Border Data Transfer’ (2017) 19 Yale Journal of Law and Technology 182, 201–4.
 See Guidelines for Virtual Currency Exchanges (n 39) s 4512N.3.
 McClarkin (n 3) 5–6.
 Burri, ‘Digital Trade Framework’ (n 58) 40.
 GATS (n 6) art XVIII. See Burri, ‘Digital Trade Framework’ (n 58) 46.
 Burri, ‘Digital Trade Framework’ (n 58) 46.
 Rolf H Weber, ‘Digital Trade in WTO-Law: Taking Stock and Looking Ahead’ (2010) 5(1) Asian Journal of WTO and International Health Law and Policy 1, 13.
 Ibid; Mitchell and Mishra (n 128) 1077–8.
 ‘Trade in Services Agreement (TiSA)’, Department of Foreign Afairs and Trade (Web Page) <https://dfat.gov.au/trade/agreements/negotiations/tisa/Pages/trade-in-services-agreement.aspx>, archived at <https://perma.cc/RST4-LF9G>; Burri, ‘Digital Trade Framework’ (n 58) 48.
 Burri, ‘Digital Trade Framework’ (n 58) 48–9.
 Ibid 51; Switzerland, Submission on Provisions on Trade-Related Principles for Information and Communication Technology Services (ICT principles) to the Plurilateral Initiative on Trade in Services, Really Good Friends: Meeting of 18 March 2013 (15 February 2013).
 Burri, ‘Digital Trade Framework’ (n 58) 52–3.
 Wunsch-Vincent, ‘Trade Rules for the Digital Age’ (n 59) 507.
 Magnus Rentzhog and Emilie Anér, ‘The New Services Era: Is GATS up to the Task?’ (Overview Paper, E15 Initiative, November 2014) 13.
 Henrys Gao, ‘Regulation of Digital Trade in US Free Trade Agreements: From Trade Regulation to Digital Regulation’ (2018) 45(1) Legal Issues of Economic Integration 47, 47 (‘Regulation of Digital Trade’). Preferential Trade Agreements (‘PTAs’) may either be free trade agreements, or customs unions with common external tariffs. In this context of regulating digital trade, and without going into a discussion of the differences in treatment of tariffs, PTAs and free trade agreements are broadly similar: see Aaditya Mattoo and Pierre Sauvé, ‘Services’ in Jean-Pierre Chauffour and Jean-Christophe Maur (eds), Preferential Trade Agreement Policies for Development: A Handbook (World Bank, 2011) 235, 258, 259.
 Wunsch-Vincent, ‘Trade Rules for the Digital Age’ (n 59) 511.
 Ibid 511–13.
 Gao, ‘Regulation of Digital Trade’ (n 233) 61.
 Wunsch-Vincent, ‘Trade Rules for the Digital Age’ (n 59) 515.
 Burri, ‘Digital Trade Framework’ (n 58) 66.
 Mitchell and Mishra (n 128) 1132.
 Wunsch-Vincent, ‘Trade Rules for the Digital Age’ (n 59) 517.
 Weber (n 225) 15.
 United States Trade Representative, TPP 14 Electronic Commerce: Chapter Summary (Report) 6 <https://ustr.gov/sites/default/files/TPP-Chapter-Summary-Electronic-Commerce.pdf>, archived at <https://perma.cc/D9PA-G4H2>; Trans-Pacific Partnership Agreement, signed 4 February 2016,  ATNIF 2 (not in force) ch 14.
 Gao, ‘Regulation of Digital Trade’ (n 233) 63–4.
 Burri, ‘Digital Trade Framework’ (n 58) 66.
 See generally Rodrigo Polanco Lazo and Sebastián Gómez Fiedler, ‘A Requiem for the Trans-Pacific Partnership: Something New, Something Old and Something Borrowed?’  MelbJlIntLaw 15; (2017) 18(2) Melbourne Journal of International Law 298; Peter Baker, ‘Trump Abandons Trans-Pacific Partnership, Obama’s Signature Trade Deal’, The New York Times (online, 23 January 2017) <https://www.nytimes.com/2017/01/23/us/politics/tpp-trump-trade-nafta.html>, archived at <https://perma.cc/5ENG-6CBA>.
 Burri, ‘Digital Trade Framework’ (n 58) 66.
 Mitchell and Mishra (n 128) 1104.
 Rentzhog and Anér (n 232) 13.
 Weber (n 225) 13.
 Sacha Wunsch-Vincent and Arno Hold, ‘Towards Coherent Rules for Digital Trade: Building on Efforts in Multilateral versus Preferential Trade Negotiations’ in Mira Burri and Thomas Cottier (eds), Trade Governance in the Digital Age: World Trade Forum (Cambridge University Press, 2012) 179, 220.
 See Chiedu Osakwe, ‘WTO Fundamentals Are Sound, but the Architecture Requires Reform and Modernisation for the 21st Century Global Economy’, E15 Initiative (Blog Post, October 2018) <http://e15initiative.org/blogs/wto-fundamentals-are-sound-but-the-architecture-requires-reform-and-modernisation-for-the-21st-century-global-economy/> , archived at <https://perma.cc/YGQ9-DVX6>.
 Burri, ‘The Governance of Data’ (n 159) 98.
 Mitchell and Mishra (n 128) 1074.
 McClarkin (n 3) 6–7.