A SINGLE DIGITAL MARKET FOR EAST AFRICA Pr s ntin vision, str t ic fr m work, impl m nt tion ro dm p nd imp ct ss ssm nt SDM EAST AFRICA A SINGLE DIGITAL MARKET FOR EAST AFRICA Presenting a vision, strategic framework, implementation roadmap, and impact assessment Casey Torgusson, Cecilia Paradi-Guilford, Isabella Hayward Ivan Gonzalez Berenguer Pena, Edgardo Sepulveda , World Bank Michael Kende, Richard Morgan, Neil Gandal, David Abecassis, Analysys Mason Standard Disclaimer Acknowledgments This volume is a product of the staff of The This report was researched and prepared by a team from the World Bank World Bank. The findings, interpretations, Group and the consulting firm Analysys Mason Limited. and conclusions expressed in this paper do not necessarily reflect the views of Analysys Mason Limited, North West Wing, Bush House, Aldwych, London the Executive Directors of The World WC2B 4PJ. Telephone: +44 (0)20 7395 9000 Bank or the governments they represent. Internet:www.analysysmason.com The World Bank does not guarantee the Registered in England No. 5177472 accuracy of the data included in this work. The boundaries, colors, denominations, The team would like to thank the following for sharing their feedback on the and other information shown on any map report, which helped strengthen the presentation and analysis: Paul Noumba in this work do not imply any judgment Um, Coordinating Country Director in Charge of Regional Integration; Boutheina on the part of The World Bank concerning Guermazi, Practice Manager for Digital Development; Tim Kelly, Lead ICT Policy the legal status of any territory or Specialist; Deo Ndikumana, Senior Operations Officer for Regional Integration; the endorsement or acceptance of Vijay Pillai, Advisor, Africa Region; Yutaka Yoshino, Program Lead and Lead such boundaries. Nothing herein shall Economist for Equitable Growth, Finance and Institutions; Yoko Doi, Senior constitute or be considered to be Financial Specialist; and Eva Clemente Miranda, Private Sector Specialist. limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. This report was prepared by the World Bank in partnership with Analysys Mason Limited, supported by a generous grant from the Swedish International Copyright © 2018 International Bank for Development Agency (Sida). Reconstruction and Development / The World Bank. Some rights reserved. World The report outlines the impetus for creating a single digital market (SDM) Bank, 1818 H Street NW, Washington, in East Africa, which would drive deeper integration and spur increased DC 20433. Telephone: +1 202-473-1000; dynamism of the digital economies of six East African countries: Burundi, Internet: www.worldbank.org Kenya, Rwanda, South Sudan, Tanzania, and Uganda. The content of this report has greatly benefited from comprehensive regional stakeholder consultation with the governments of the six countries, as well as the private sector, civil society organizations, development partners, and academia. Multi-stakeholder workshops took place in Kenya, Rwanda, Tanzania, and SDM EAST AFRICA Uganda between April and May 2017. The research team has also gathered inputs through a series of bilateral meetings and a qualitative online survey. These consultations, along with desk research, have helped shape a joint vision and strategic framework for the SDM presented in this report, including anEAST SDM AFRICA assessment of the state of existing markets. B A SINGLE DIGITAL MARKET FOR EAST AFRICA Content Foreword iv Figures Executive Summary 1 1. The building blocks of the Figure 1: Overview of SDM vision and Single Digital Market 6 strategic framework 7 1.1 Leveraging the lessons learned from Figure 2: Population size of the largest global SDM initiatives 6 global markets 16 Figure 3: Internet users and broadband 1.2 A vision for a Single Digital Market in East Africa 6 penetration rates, 2017 18 1.3 The SDM strategic framework 7 Figure 4: 2G, 3G, and 4G mobile coverage, based on population, 2017 19 1.4 Action andcooperation - towarda ‘Digital Roadmap’ 11 Figure 6: Mobile broadband prices as % of GNI per capita 22 2. Impetus for the Single Digital Market 14 Figure 7: Date of entry into force for a 2.1 Economic and jobs impact 14 selection of agreements and treaties regulating the international 2.2 Drivers of growth and job creation 15 protection of IP 24 3. Assessment of current digital markets 18 Figure 8: Share of population that is unregistered 26 3.1 A single connectivity market 18 Figure 9: Penetration of financial institution 3.2 A single data market 23 accounts versus mobile money accounts 27 3.3 Single online market 26 Figure 10: Mobile payment systems 3.4 Key enablers 32 deployed in the EAC 28 Figure 11: E-Government Development 4. Conclusions and next steps toward Index scores 30 implementation 38 Figure 12: Good country scores 30 Annex A: SDM Roadmap 39 Figure 13: Access to electricity, 2016 32 Annex B: SDM Scorecard 45 Figure 14: Ease of doing business index, 2016–2017 33 Annex C: Existing integration efforts 53 Figure 15: Tech hubs and incubators Annex D: Economic Impact Assessment 58 by country 35 Figure 16: Adult literacy rate, 2015 36 A SINGLE DIGITAL MARKET FOR EAST AFRICA iii Foreword East Africa cannot afford to think small. At the current, incremental pace of economic and social advancement, too many of today’s youth will continue to be denied the opportunity to live up to their potential. The rise of digital technologies offers a chance to disrupt this trajectory, unlocking new pathways for rapid economic growth, innovation, job creation, and access to services which would have been unimaginable only a decade ago. The burgeoning tech start-up clusters in Nairobi, Dar, Kampala, and Kigali provide inspiration for what the future could hold. Tapping into this potential will require bold, visionary leadership and deeper integration of the region’s digital economy and innovation ecosystem. By working together and seizing opportunities to ‘leapfrog’ outdated infrastructure, technology, and business models, East African countries can position the region as a premier digital investment and innovation destination. In isolation, East African countries will miss out on this opportunity, left behind by rapid technological advancement and rivals with large domestic or integrated regional markets and more proactive digital investment and reform strategies. To become tomorrow’s innovators, entrepreneurs, and leaders, East Africa’s youth need to be empowered with the digital skills and access to technology, ideas, and markets that are essential to thrive in an increasingly digitized global economy. Governments need to find more nimble and effective means of delivering services and interacting with citizens. Businesses need to utilize digitally centered business models to connect with the millions of customers previously out of reach due to geography or low incomes. East African countries simply do not have the size and resources to meet all these needs on their own nor the luxury of time to address these challenges gradually without falling further behind. The East Africa Single Digital Market (SDM) initiative is designed to help the region’s citizens, businesses, and governments rise to the challenge and seize opportunities at hand. It aims to bring together the region’s digital leaders and stakeholders to rally around a common vision for the region’s digital aspirations and a coordinated blueprint that can turn this vision into reality. By working together, East African countries and their development partners can maximize the impact of their investments and reforms to enhance their economic competitiveness and achieve digital transformation far greater than the sum of its parts. iv A SINGLE DIGITAL MARKET FOR EAST AFRICA Executive Summary Objectives and Definition of An SDM for East Africa is defined as one in which cross-border barriers to providing and accessing digital infrastructure, a Single Digital Market content, and services are eliminated. It is one in which a seamless and competitive regional digital ecosystem drives a reinforcing The East Africa Single Digital Market (SDM) initiative aims to cycle of economic growth, investment, innovation, job creation, support the region to become a more deeply integrated and and improved service delivery. dynamic digital investment, innovation, and growth hub, building on the underlying strengths of the domestic digital economies and the ongoing political and economic integration SDM Framework processes in East Africa as well as at the continental level. It Achieving an SDM will require simultaneously supporting domestic stems from a recognition that countries in the region are too development and cross-border integration of the following small to succeed in the digital economy in isolation. They need submarket structures, which form distinct yet interconnected the economies of scale and network effects offered by a larger layers of the overall SDM. These include the following: regional market to help bridge the digital divide faster within their respective countries and the region in relation to the global digital • A single connectivity market, which will remove barriers economy. An SDM could accelerate the growth of technology- to regional telecom infrastructure and services deployment enabled businesses, lower the cost of key telecom services, and to encourage investment, improve performance, eliminate catalyze new digital services for citizens and businesses alike. pricing and quality differentials between coastal and landlocked countries while simultaneously expanding access An integrated East African digital market would be the ninth to connectivity to all. largest in the world, based on population, creating the ‘domestic’ • A single data market, which will enable secure exchange, market size needed to attract digital investment and provide storage, and processing of data across borders; support a larger and more familiar space for local firms to grow before regional deployment of data infrastructure; and drive supply having to compete in global markets. It will provide the region with and demand for data-driven services and innovation across the heft to compete against digital giants on the continent such the region. as Nigeria, and ultimately to contend with the globally dominant • A single online market, which will allow firms, governments, digital hubs such as Silicon Valley, which benefit from large and citizens to access and deliver both public and private domestic markets in which to scale before tackling other markets. services online; undertake e-commerce transactions; and access digital content and information seamlessly from The initiative has brought together stakeholders from the public anywhere in the region. and private sectors, civil society, and academia to highlight the Advancement in each distinct market layer is expected to create a aspirations for, and benefits of, creating an SDM, as well as to virtuous cycle. Each layer builds on the other, which will reinforce articulate: the development, expansion, and integration of the SDM and • A shared vision for the SDM, further drive access to the internet, innovation, job creation, and growth. • A strategic framework and roadmap for achieving the SDM, and Joint action will also be required to foster a conducive regional • A market assessment and scorecard that captures the enabling environment to underpin the creation and impact of current state of domestic markets and regional readiness an SDM, including targeted support for digital skills development, to move toward an SDM, as well as targets and metrics for digital innovation and entrepreneurship, improvements to tracking progress. regional logistics and supporting infrastructure such as energy The SDM initiative for East Africa draws on the successes and and transport, and investment promotion and access to capital. lessons learned from similar initiatives implemented elsewhere, Finally, strong leadership and institutions, supported by effective which includes the European Digital Single Market. Yet, it is coordination, communication, capacity building, and data adapted to the unique context, challenges, and opportunities collection, will be necessary to spearhead the SDM initiative of the region and within each country. moving forward. A SINGLE DIGITAL MARKET FOR EAST AFRICA 1 Current Market Assessment potential. Most online services originate from outside the region, dominated by large firms that have the resources to navigate the The comprehensive market analysis and readiness assessment web of procedures, licenses, taxation, and other barriers which carried out for this study identified key barriers that will need to require businesses to set up relatively independent operations in be addressed to achieve an SDM, as summarized below: each country rather than seamlessly expanding across borders. The region is a world leader in digital financial services, but the Despite gains to date, increasing the affordability, availability, and lack of interoperability and high fees for cross-border digital adoption of high-speed internet and digital technologies remain transactions discourages regional e-commerce. a major challenge. While mobile voice services have grown rapidly, internet access is still very low in most parts of East Africa. Key digital economy enablers such as digital literacy and more Landlocked countries are still paying higher premiums to access advanced, workforce-ready skills, logistics infrastructure and bandwidth of lower quality, compared to their coastal neighbors. services, and access to energy are quickly improving but still severely constrain demand for digital services and productivity The region’s data infrastructure remains underdeveloped, in part of firms. Uncoordinated taxation policy and high tax burdens due to lack of a clear legal and regulatory regime to support across all layers of the digital value chain are likewise raising free flow, storage, and processing of data across borders and costs and discouraging investment. Support for innovation and a harmonized data protection and privacy regime to ensure entrepreneurship are helping generate promising new ideas and security of personal data. The availability and standardization of start-ups but these firms still struggle to access capital and scale. public data sets to drive development of data-driven services and analytics is improving but insufficient. Cybersecurity collaboration is improving but is in early stages. Digital Roadmap To address these constraints and harness the region’s potential, Innovative online services and content are emerging across a ‘Digital Roadmap’ of priority policy reforms and investments the region, but barriers to cross-border expansion and access has been identified, presented in full in Annex A and with a by users and impediments to investment are holding back summary of the top priority actions presented below: Digital Roadmap Summary: Top Priority Actions Action i. Undertake coordinated public-private partnership (PPP) investments toward achieving universal broadband access across the region (regional backbones and access networks/last mile) ii. Undertake coordinated reduction in taxation and fees to stimulate investment and improve affordability and access to communications services • Reduce or eliminate import duties for network equipment and digital devices SINGLE • Reduce fees and facilitate easier access to rights-of-way to lay cables and mount equipment along other CONNECTIVITY public infrastructure (roadways, electricity transmission lines and streetlight poles, pipelines, rail, and so on) MARKET • Lower cost or free spectrum allocation for service provision in rural areas • Reduce or eliminate value added tax (VAT) for telecommunication services iii. Extend One Network Area (ONA) coverage to more countries and services • Expand ONA for voice to Tanzania • Expand ONA for data to Burundi, Tanzania, and South Sudan iv. Establish a regional backbone interconnection regime based on open access principles with rights for any licensed operator in the region to purchase wholesale capacity/access submarine cables and a glidepath for lowering interconnection rates 2 A SINGLE DIGITAL MARKET FOR EAST AFRICA Action i. Remove undue legal and regulatory restrictions (for example, data nationalization) on the free flow, storage, and processing of data across borders ii. Harmonize data protection, privacy, and data exchange laws and regulations with clear guidelines SINGLE DATA for cross-border data flows, personal and sensitive data, and accounting for emerging services MARKET (cloud services, data analytics, and so on) iii. Digitize key government registries using regional standards to enable cross-border interoperability and implement a regional open data initiative using shared data standards to make regional data sets available for public and private sector data-based services and analytics iv. Establish a regional cybersecurity task force Action i. Establish mutual recognition of national digital IDs and a regional platform for identification SINGLE ONLINE verification by governments and digital services providers MARKET ii. Establish full interoperability between mobile money networks (domestically and regionally) with a glidepath toward lower exchange fees for cross platform and cross-border transactions iii. Adopt harmonized e-transactions and consumer protection laws, based on the East Africa Community (EAC) Electronic Transactions Act Action i. Establish Regional Centers of Excellence for advanced information and communication technology (ICT) education and research ii. Expand rapid technology skills training, such as coding boot camps and support to mass digital ENABLING literacy programs REGIONAL iii. Develop the ‘Digital East Africa’ brand and enhance regional innovation and entrepreneurship ENVIRONMENT support networks iv. Develop regional investment promotion strategy and marketing campaign v. Facilitate ‘upskilling’ and networking of national and regional tech hubs, incubators, and accelerators vi. Cultivate a regional pipeline of start-ups with demonstrated proof of concept ready to attract international venture capital/angel investors vii. Establish a regional tech entrepreneur association to facilitate government-industry dialogue A SINGLE DIGITAL MARKET FOR EAST AFRICA 3 Economic Impact Assessment aspirations. Notably, the economic analysis suggests that those countries currently lagging in connectivity will enjoy the most The impact on gross domestic product (GDP) growth, job significant boost from an integrated digital market. creation, and poverty reduction is also expected to be significant: implementing an SDM is estimated to generate up to a US$2.6 Next Steps billion boost in GDP and 4.5 million new jobs and strongly benefit those at the bottom of the pyramid. This report serves as a first step in the SDM initiative, helping concerned parties to rally around a joint vision and overarching These estimates are conservative as they do not fully take framework. The intention is now to identify programs and into account the spillover effects from a more integrated and resources to support countries in the region to implement the competitive regional market, which is expected to increase priority actions identified as part of the Digital Roadmap. innovation, technology adoption, and investment across all sectors over the medium to long term, fueling a reinforcing cycle of By taking proactive action and working together rather than in productivity gains, growth, and job creation for many years to come. competition, the SDM can be a win-win for all countries in East Africa. Actions proposed will capitalize on the momentum of While the countries of East Africa feature differing levels of digital ongoing regional integration initiatives, such as the successful maturity, all stand to gain significantly from an SDM. Firms and ONA roaming agreement, while unlocking opportunities for innovators will gain greater access to the regional market and accelerating the wider East African integration agenda. The a larger pool of online consumers and venture capital. Citizens initiative will leverage and work through existing regional will gain access to more relevant and lower-cost digital content institutions and platforms, such as the East Africa Communications and services and the skills and opportunities to participate and Organization (EACO), EAC, Northern Corridor Committee, and at thrive in the new economy. Governments will be better equipped the continental level, through the African Continental Free Trade to meet their economic development and service delivery Area and Smart Africa Alliance. Figure ES.1. Key components of the SDM report SDM VISION CURRENT STATE ASSESSMENT DIGITAL ROADMAP SINGLE ONLINE MARKET SOUTH SUDAN POLICY TOOLS SINGLE DATA MARKET UGANDA KENYA SINGLE CONNECTIVITY RWANDA INVESTMENT MARKET TANZANIA BURUNDI ENABLING ENVIROMENT TARGET INDICATORS CURRENT STATE ASSESSMENT ACTIONS TO IMPACT INDICATORS SDM SCORECARD 4 A SINGLE DIGITAL MARKET FOR EAST AFRICA Introduction Rising mobile phone penetration, improving broadband Internet connectivity, and widespread adoption of mobile money across “ East Africa are changing the way the region communicates, SDM is about creating collaborates, and transacts. This digital evolution has spurred development of a small, but rapidly growing tech sector, economies of scale particularly in urban hubs, with innovative entrepreneurs launching new digitally enabled services and creating 21st and capitalizing century jobs. More significantly, adoption of digital technology is gradually driving productivity gains in traditional brick and mortar on the network industries. Governments across the region have likewise begun to effects necessary to take advantage of these trends by moving public services online, utilizing data to improve policy making and digital platforms to increase the region’s increase the efficiency of public service administration. ” competitive edge. While this digital evolution shows significant promise, it is not yet amounting to a revolution. Access to the internet is still very low by international standards. Internet user figures, reported by the International Telecommunication Union (ITU), suggest that only 17 percent of people in the region were online in 2017.1 Though it is improving, internet service quality remains poor in most countries and unaffordable to large parts of the population. For East Africa to benefit from the transformational social and economic benefits generated by digital technology and increasing connectivity, or so-called ‘digital dividends’, much more needs to Building on its relatively strong trade, financial, and political be done to bridge this ‘digital divide.’ integration protocols and institutions, as well as harnessing the dynamism and reputation of its growing tech clusters, deeper Domestic digital markets are small and fragmented and of cross-border integration of digital markets and closer coordination insufficient size to attract investment or provide the addressable between East African countries could drive the creation of a customer base for technology-driven companies to rapidly scale. strong and competitive regional digital ecosystem that is able to Basic digital literacy levels remain low and the region’s educational attract investment, promote innovation, and spur job creation. institutions are not producing the volume of workforce-ready computer and network engineers, software developers, and Ultimately, the drive for an SDM is about creating economies other high-skilled digital professionals needed, subsequently of scale and capitalizing on the network effects necessary constraining growth and investment by tech and telecom firms to increase the region’s competitive edge. This is particularly in the region. Governments are only scratching the surface of the important in the context of the digital economy, given the high opportunities associated with leveraging technology to improve fixed costs of telecom infrastructure and data centers, the need service delivery. for tech start-ups to rapidly scale and capture market share, and the exponential growth in value of digital platforms and services While these challenges are not unique to East Africa, the region is as increasing numbers of users adopt the services and generate uniquely placed to overcome them and to thrive as a leading hub in more data and content. the global digital economy. As the birthplace of some of the most innovative digital solutions in Africa, including mobile money, the This report aims to outline the case for the creation of an SDM, region is well positioned to capitalize on the momentum created define the key elements that comprise an SDM, and inform the to date in moving toward a single digital market (SDM). approach toward its implementation. 1 ITU World Telecommunication/information and communication technology (ICT) Indicators database. Regional average based on country-level internet penetration weighted by country population. A SINGLE DIGITAL MARKET FOR EAST AFRICA 5 1. The building blocks of the Single Digital Market 1.1 Leveraging the lessons learned from “ A single digital market is one in global SDM initiatives which cross-border barriers to The SDM initiative for East Africa, proposed in this report, draws providing and accessing digital on the successes and lessons learned from similar initiatives infrastructure, content, and implemented elsewhere.2 The largest established example to date is the European Digital Single Market, which allows individuals services are eliminated and one and businesses to “seamlessly access and exercise online in which a seamless, competitive activities under conditions of fair competition, and a high level of consumer and personal data protection, irrespective of their regional digital ecosystem nationality or place of residence.” It is estimated that accelerating drives a reinforcing cycle of Europe’s digital potential, through a single market initiative, could economic growth, investment, add trillions of euros to economic growth in less than a decade.3 innovation, job creation, and ” The European Digital Single Market strategy comprises a wide improved service delivery. range of initiatives that help achieve this vision. These include, but are not limited to, (a) achieving universal availability of superfast broadband (30 Megabits per second (Mbps)) across all member Africa. The vision, strategic framework, and roadmap adopted for states by 2020; (b) eliminating roaming prices; (c) harmonizing achieving an SDM in East Africa thus needs to be anchored in the spectrum licensing regimes; (d) ensuring the free flow of data unique challenges and strengths that characterize regional and across the region; (e) cooperating on issues such as cybersecurity, domestic markets. An assessment of these market structures is privacy, and data protection; (f) addressing value added tax presented in Chapter 3 of this report. However, the overarching (VAT) and geo-blocking issues to make cross-border e-commerce strategic framework presented here is likely to be applicable to easier; and (g) harmonizing laws in areas such as copyright, online similar regional integration elsewhere in Africa, including the contracts, and consumer protection. recently launched pan-African initiative, under Smart Africa, to move toward a Single African Digital Market (see Annex C for more details). 1.2 A vision for a Single Comprehensive stakeholder consultations revealed a diversity of Digital Market in East Africa views of what an SDM would mean for East Africa, before ultimately coalescing around the following vision statement (the ‘SDM Vision’): While there is value in drawing on other SDM initiatives, all regions have different market structures and characteristics. Solutions The vision articulated is grounded in the region’s aspirations to that work elsewhere may therefore not be entirely suitable in East forge a new path for East Africa’s socioeconomic development, pairing an ambitious regional integration agenda with the 2 Nascent SDM initiatives have also begun in Latin America and Eurasia with similar aims of digital market integration. desire to unlock the opportunities presented by the spread and 3 See https://www.mckinsey.com/business-functions/digital-mckinsey/our-insights/ development of innovative digital technologies and the growth of digital-europe-realizing-the-continents-potential. the wider global digital economy. 6 A SINGLE DIGITAL MARKET FOR EAST AFRICA 1.3 The SDM strategic framework The SDM framework consists of three markets or layers and an Ultimately, each layer of the SDM builds on the other. However, enabling environment foundation, as detailed in Figure 1. A vibrant with the development of each distinct layer a positive feedback SDM will be based on promoting the development of all three loop will emerge, which reinforces the expansion and integration markets and the enabling foundation within each country while of the SDM. For example, while connectivity is a prerequisite for simultaneously removing cross-border barriers to integration the development of online services, as the online services market (regionally and in some cases globally) within each layer to create expands so will the demand for connectivity infrastructure, as a seamless regional market. access to the internet becomes more attractive to new users. Figure 1: Overview of SDM vision and strategic framework PROMOTE THE DIGITAL MARKET REMOVE CROSS-BORDER BARRIERS Digital ID Digital payments Ensure e-commerce, digital services E-transactions and the functions that support them Consumer protection all work across borders Digital public services SINGLE Remove trade and customs barriers ONLINE for goods purchased online Trade and customs MARKET Logistics Ensure the data protection and Data protection and privacy privacy laws allow for cross-border Cybersecurity data transfers Content regulation SINGLE DATA Share cybersecurity resources in MARKET the region Infrastructure Remove cross-border barriers to infrastructure and connectivity Services (wholesale and retail) SINGLE CONNECTIVITY MARKET ENVIRONMENT HARD INFRA’ LEADERSHIP (E.G.POWER) INNOVATION Where relevant, create FINANCING scale for these ENABLING CAPITAL DIGITAL DIGITAL SKILLS analogue complements ENVIRONMENT across the region A SINGLE DIGITAL MARKET FOR EAST AFRICA 7 A single connectivity market the landlocked countries in the region, which currently face high transmission charges to access international connectivity through To participate in the digital economy, firms, citizens, and their coastal neighbors. governments must first be online. A single connectivity market At a wholesale level, this could be achieved by ensuring interconnection is one in which the capacity, speed, and quality of connectivity and interoperability of national backbone networks across the region services needed to support the free flow of data and online (physical and regulatory)—facilitating the construction of new cross- service provision is available from anywhere in East Africa at border networks and streamlining licensing requirements or obligations an affordable price. This would apply to both the retail-level for regional players. At a retail level, this could mean extending existing connectivity services available to end users, as well as wholesale regional roaming initiatives, under One Network Area (ONA), to cover connectivity made available to the internet services providers data (see Focus Box 1 for more details on the ONA). (ISPs) that serve them. The economies of scale created by a single connectivity market would Cross-border barriers to both connectivity infrastructure and also help improve the affordability of related services and infrastructure services deployment and affordability would be removed, deployment. By lowering wholesale and retail costs and increasing the creating a larger, single connectivity market, with equal access to availability of connectivity infrastructure across the region, a single infrastructure (for example, network coverage), as well as a level connectivity market would help bring more people online, bridging playing field for wholesale connectivity services (for example, the current ‘digital divide’ both within and between countries in equivalent pricing for international internet protocol (IP)), the region. transit and retail services (for example, reducing or eliminating roaming rates). Removing these barriers would help infuse Increasing connectivity, through a single connectivity market, is a greater competition, which would in turn help drive down prices. necessary stepping stone for the development of a single data market. It would also minimize differences in international capacity and The only way that data can flow freely across borders is on top of prices between countries. A single connectivity market would seamless regional networks. Equally, a larger online population across thus have limited or no price or performance differentials for the region is critical in making the development of a vibrant single data connectivity across the region. This feature would be critical for and online market viable. Focus Box 1 ONA - A successful case of regional digital market integration High roaming charges and interconnection/termination rates can The impact of lower tariffs on individual roamers has been create significant barriers that stifle cross-border communications. significant, with cross-border voice traffic growing rapidly. In 2014, the countries of the East African Community thus made In Kenya and Uganda, cross-border voice traffic has tripled. a joint commitment to fast-track the creation of an ONA. The ONA Meanwhile, Rwanda and South Sudan both experienced a fivefold currently covers Kenya, Rwanda, Uganda, and South Sudan through and threefold increase, respectively, in the wake of the ONA’s an accelerated integration program under the Northern Corridor introduction. As roaming represents a minor source of revenue Initiative, with the possibility of Tanzania and Burundi joining when for mobile network operators (MNOs), the initiative is not seen to they are ready. The agreement introduced harmonized rate caps for have had an adverse impact on industry. cross-border traffic originating and terminating within participating The ONA is an example of what can be achieved when target ONA countries and elimination of roaming surcharges for users initiatives are prioritized and fast-tracked at the regional level but travelling within the region. The ONA was initially applied to voice also the price sensitivity of consumers—suggesting that other services in all four countries and has more recently been extended interventions that help lower connectivity prices for consumers to data in Kenya, Uganda, and Rwanda, though reports suggest the will stimulate growth in the regional uptake of digital and latter is not yet fully complete. information and communication technology. In other words, fully In Uganda, retail roaming rates were cut from US$0.93 to US$0.10 extending the initiative to data would likely help increase cross- per minute (based on figures from 2016) following the introduction border data exchange in the region. of the ONA. Source: Kelly, T., and C. Kemei. 2016. WDR 2016: Digital Dividends; A Case Study of ONA. ITU.4 4 See http://pubdocs.worldbank.org/en/499731452529894303/WDR16-BN-One-Network-Area-in-East-Africa-Kelly-Kemei.pdf and https://www.itu.int/dms_pub/itu-d/opb/ pref/D-PREF-EF.ONA-2016-PDF-E.pdf. 8 A SINGLE DIGITAL MARKET FOR EAST AFRICA Single data market “ The digital economy is driven by data. Globally, cross-border A single data market would data flows have increased some 45 times since 2005. In the past unlock the data-driven two years alone, this cross-border data exchange is estimated to economy in East Africa, have boosted global gross domestic product (GDP) by 10 percent and is worth some US$7.8 trillion.5 This trend shows no signs of allowing data to flow abating.6 According to IBM, 90 percent of all data that exists today freely and securely across have been created within the past 12 months, growing at a rate of borders—to be processed, 2.5 quintillion bytes per day.7 In sum, data information flows now have a more significant impact on GDP growth for countries that analyzed, stored, or used anywhere ” trade in traditional goods.8 Looking ahead, more and more critical infrastructure and new services will be heavily data driven. This includes the Internet of Things (IoT), which will involve adding connectivity to many common devices we use every day, as well as critical infrastructure such as power networks. This offers However, robust data protection and privacy laws, as well as great potential but needs to be both proactively nurtured as a cybersecurity procedures would need to be in place to protect source of growth and productivity while also carefully managed sensitive data, maintain consumer confidence in data-driven to mitigate risks. services, and facilitate data sharing of governmental records. For example, harmonized regulatory frameworks would need A single data market would unlock the data-driven economy in to be established, as well as joint platforms to boost regional East Africa, allowing data to flow freely and securely across readiness to prevent and address shared cybersecurity threats borders—to be processed, analyzed, stored, or used anywhere and related attacks. in the region, subject to compliance with appropriate regulation. This would require the harmonization of data laws and the creation In an SDM, there would be no undue restriction on content of explicit terms under which data could be safely stored and coming into a country from elsewhere in the region, thus exchanged between countries (similar to existing European ‘Safe precluding practices such as geo-blocking. Equally, intellectual Harbor’ rules), in turn supporting regional deployment and access property (IP) rights and content protection would need to be to data infrastructure and data-driven services and innovation. coordinated and enforced. An East African single data market would encourage the creation The creation of a larger market would also generate substantial of a much larger pool of data, which could enable data-driven cost savings by creating economies of scale that make investment innovation (DDI) and ‘big data’ analytics, resulting in significant in regional data centers that support online services, including economic and social benefits and efficacy gains across virtually all cloud hosting, more financially viable. However, achieving this sectors. Governments would support the availability of open data would necessitate the removal of data localization requirements from public, and potentially also private, data sets, by adopting that confine data storage within national borders. regionally (and globally) standardized, machine-readable formats Ensuring the free flow of data is an essential enabler for many that could foster the development of DDI. online services, such as e-commerce, digital payments, and other cloud services, to work across borders. These services are built 5 ‘Single Digital Market for Africa Report’, Transform Africa Summit May 2018. on large-scale data. Free-flowing data is also a prerequisite for 6 See Data-Driven Development. 2018 (forthcoming). Information and Communication for Development. World Bank. any regional cooperation on digital ID, which is essential for many 7 IBM 2016, see https://www-01.ibm.com/common/ssi/cgi-bin/ online transactions. A single data market is therefore a critical ssialias?htmlfid=WRL12345USEN. first step to encouraging the creation of new and innovative data- 8 See https://www.mckinsey.com/business-functions/digital-mckinsey/our- insights/digital-globalization-the-new-era-of-global-flows. driven solutions and a vibrant single online market. A SINGLE DIGITAL MARKET FOR EAST AFRICA 9 Single online market Regional enabling In a single online market, innovators, entrepreneurs, and environment regional companies would offer digitally enabled services and content freely across the region, with no barriers to cross- The three market layers would need to be underpinned by border delivery and access of digital goods or services. For a supportive enabling environment at both the national and countries in the region with more developed markets for online regional level, which would help promote the development services, such as e-commerce, this would stimulate growth by of the regional digital economy. A vibrant digital ecosystem enabling their local businesses and entrepreneurs to reach requires the development of digital skills, a favorable business a much larger regional market. In all countries in the region, and innovation environment, deployment of hard infrastructure, citizens and businesses would gain larger access to a wider access to capital financing, and effective digital leadership and range of digitally enabled services. It would, for example, institutions. Collaboration between countries can complement increase the reach, efficiency, and quality of public services, and significantly enhance efforts to improve the enabling by supporting the digitization of government. E-government environment at the national level. services would also help familiarize consumers with the use Digital literacy and skills would need to be improved to help of online services and increase the consumer base for these users have meaningful access to the internet, digital information, services. Regional coordination of related initiatives in areas and services (both public and private), as well as create content such as open data would also help stimulate the development and even digital businesses. The promotion of advanced digital of regional services and content. skills would support the development of the digital economy, by A single online market would involve cross-border collaboration training the innovators and entrepreneurs who would develop the on a series of prerequisite technologies and legal frameworks, next generation of digital services and to fill technology-related including (a) digital identification (ID), (b) digital payments, jobs opening up across all industries. Regional collaboration can (c) e-transaction legislation, and (d) consumer protection. help create the scale for investments in centers of excellence to Customers must be able to go online, identify themselves (if develop more advanced skill sets demanded by industry and to needed), sign and pay for goods and services digitally, and feel support regional innovation. confident that their personal data and consumer interests are An enabling environment would make it easier to start new being protected. digital companies, accelerate their growth domestically and Certain online services, such as finance, health, and travel, regionally, and support innovation and technology adoption will require a digital means of identification, which would among established businesses. A more supportive domestic need to be recognized across borders for a regional market to and regional business and innovation environment would need emerge. Digital payment platforms would need to be available, to be encouraged, which could support the emergence of new affordable, and interoperable across the region and work and innovative online service providers, help entrepreneurs globally, enabling users in East Africa to pay for international to develop digital content and services that are locally and services. Harmonized e-transactions legislation would also regionally relevant, while creating new jobs in the region. Better need to be in place to create legal equivalence between paper- access to capital as well as the promotion and facilitation of based and electronic transactions, enabling digital signatures. investment would provide support for both start-ups (such as new online service providers) and larger established companies For service providers, including government agencies and (such as infrastructure providers). A larger portfolio of potential companies, a single online market would enable the efficient investments and networks of tech incubators and accelerators delivery of goods and services purchased and/or delivered across the region, as well as proactive regional branding and online. This would be achieved by a more harmonized investment promotion would also help make the East African approach to trade and customs arrangements, lower tariffs, market more attractive to potential investors and venture capital. and the elimination of nontariff barriers. 10 A SINGLE DIGITAL MARKET FOR EAST AFRICA The technology that drives the SDM requires affordable momentum of ongoing regional integration initiatives such and reliable access to power—be it as a means of allowing as the ONA, the East African Community (EAC) Common the average consumer to charge their digital devices or Market, and the African Continental Free Trade Area (AfCFTA). for businesses and governments to provide reliable digital Implementing the recommendations will require coordinated services and operate communications networks. Equally, solid efforts between national governments, in collaboration transportation infrastructure would be required to facilitate with regional bodies such as the EAC, the East Africa the logistics services needed to support e-commerce. While Communications Organization (EACO), Northern Corridor both power and transport infrastructure are critical enablers of Integration Projects (NCIPs), and the Smart Africa Alliance. an SDM, these deficits would largely need to be tackled outside Support from private sector organizations, nongovernmental of the SDM roadmap, through parallel investment programs, organizations (NGOs), academia, and donor partners will coordinated at the regional level. However, an SDM can also likewise be critical to the design and implementation of support innovative means of improving infrastructure and policy reforms and investment programs. A summary of logistics services. These include off-grid solar provision enabled ongoing integration efforts, including work undertaken by by mobile payment technology, improved logistics through pertinent regional bodies and fora that are critical to the digital mapping and cargo drone delivery, and facilitation of implementation of the SDM is summarized in Annex C. cross-border trade through regionally integrated customs and immigration information systems and regional data exchange In addition to working with said institutions and stakeholders, and privacy protocols. coordinating these parallel efforts will be necessary. This will ensure clarity on roles and responsibilities, avoid duplication, Finally, effective digital leadership and institutions will need to support capacity building, bridge existing data gaps, and allow be in place at both the national and regional level, and be able for effective monitoring of progress. Consultations suggest to coordinate and harmonize relevant policy, regulation, and that a designated coordination body might be necessary to investments, in line with the overall SDM strategic framework facilitate this process. This body could also help communicate and roadmap. the vision articulated by the SDM initiative. 1.4 Action and By taking proactive action and working together rather than in competition, the result can be a win-win for all East African cooperation - toward countries. Most importantly, it can help close the ‘digital a ‘Digital Roadmap’ divide’ among East Africa’s people, ensuring that everyone has access to and is benefiting from digital technologies Based on the vision and strategic framework (detailed earlier), rather than just a privileged few. as well as a comprehensive analysis of the readiness of domestic Existing SDM strategies employ scorecards to track and regional markets (detailed in Chapter 3), a series of priority implementation of the overarching strategic framework action areas have been identified. These will need to be over time. The European strategy is, for example, linked to addressed at both the regional and national level for an SDM measurable targets. It defines several quantifiable indicators to emerge. These action areas are presented here as part of a against which progress can be tracked in each country. consolidated SDM roadmap. A summary of priority actions under Indicators are regularly published through the Digital the proposed roadmap is presented in Focus Box 2, whereas the Scoreboard and the Digital Economy and Society Index. The full version can be found in Annex A. Priority actions have been scorecard proposed for the SDM in East Africa, detailed in selected based on expected impact and feasibility. Annex B, replicates this model but with indicators and targets While many actions identified in the roadmap are ambitious, selected to correspond to the unique East African context none are unachievable. The roadmap hopes to capitalize on the and aspirations. A SINGLE DIGITAL MARKET FOR EAST AFRICA 11 Focus Box 2 Digital Roadmap: Top priority actions9 Action i. Undertake coordinated PPP investments toward achieving universal broadband access across the region (regional backbones and access networks/last mile) ii. Undertake coordinated reduction in taxation and fees to stimulate investment and improve affordability and access to communications services • Reduce or eliminate import duties for network equipment and digital devices • Reduce fees and facilitate easier access to rights-of-way to lay cables and mount equipment along other public infrastructure (roadways, electricity transmission lines and streetlight poles, pipelines, SINGLE rail, and so on) CONNECTIVITY • Lower cost or free spectrum allocation for service provision in rural areas MARKET • Reduce or eliminate VAT for telecommunication services iii. Extend ONA coverage to more countries and services • Expand ONA for voice to Tanzania • Expand ONA for data to Burundi, Tanzania, and South Sudan iv. Establish a regional backbone interconnection regime based on open access principles with rights for any licensed operator in the region to purchase wholesale capacity/access submarine cables and a glidepath for lowering interconnection rates Action i. Remove undue legal and regulatory restrictions (for example, data nationalization) on the free flow, storage, and processing of data across borders ii. Harmonize data protection, privacy, and data exchange laws and regulations with clear SINGLE DATA guidelines for cross-border data flows, personal, and sensitive data and accounting for MARKET emerging services (cloud services, data analytics, and so on) iii. Digitize key government registries using regional standards to enable cross-border interoperability and implement a regional open data initiative using shared data standards to make regional data sets available for public and private sector data-based services and analytics iv. Establish a regional cybersecurity task force 9 This table serves as a selection of the top priorities from among the full Digital Roadmap recommendations found in Annex A. 12 A SINGLE DIGITAL MARKET FOR EAST AFRICA Action i. Establish mutual recognition of national digital IDs and a regional platform for identification SINGLE verification by governments and digital services providers ONLINE ii. Establish full interoperability between mobile money networks (domestically and regionally) MARKET with a glidepath toward lower exchange fees for cross-platform and cross-border transactions iii. Adopt harmonized e-transactions and consumer protection laws, based on EAC Electronic Transactions Act Action i. Establish Regional Centers of Excellence for advanced ICT education and research ii. Expand rapid technology skills training, such as coding boot camps and support to mass digital literacy programs ENABLING iii. Develop the ‘Digital East Africa’ brand and enhance regional innovation and entrepreneurship REGIONAL support networks ENVIRONMENT • Develop regional investment promotion strategy and marketing campaign • Facilitate ‘upskilling’ and networking of national and regional tech hubs, incubators, and accelerators • Cultivate a regional pipeline of start-ups with demonstrated proof of concept ready to attract international venture capital/angel investors • Establish a regional tech entrepreneur association to facilitate government-industry dialogue A SINGLE DIGITAL MARKET FOR EAST AFRICA 13 2. Impetus for the Single Digital Market 2.1 Economic and jobs impact “ Generating an An East Africa SDM will enable the region to capture a bigger share of the global digital economy, driving economic growth and diversification, creating up to US$2.6 jobs, and improving services. In 2016, the digital economy was worth US$11.5 trillion, or 15.5 percent of global GDP. It is expected to reach 25 percent in billion boost in less than a decade, far outpacing the growth of the ‘traditional’ economy. regional GDP East African countries are capturing only a tiny fraction of these benefits. An SDM offers the opportunity for the region to disrupt its growth trajectory by and 4.5 million providing its businesses a more solid footing to compete and empowering its citizens to thrive in the global digital economy. Without deeper integration and ” new jobs cooperation, East African countries will largely continue to miss out on this ‘4th industrial revolution’, or ‘Industry 4.0’, with only a small elite benefiting while the rest are left behind. An economic assessment undertaken to estimate the impact of implementing an East Africa SDM conservatively projects a boost to regional GDP of between US$0.93 and US$2.6 billion over five years (between 0.57 percent and 1.60 percent additional growth), driven at first by increased access to the internet, with accelerating growth thereafter driven by development of new digital industries and services and adoption of technology and digital platforms by traditional firms. It further projects creation of between 1.6 million and 4.5 million new jobs (between 2.2 percent and 6.2 percent additional growth) over the same period. The analysis also predicts that the impacts will be felt across the income spectrum, with those at the bottom of the pyramid gaining disproportionately as citizens and businesses previously shut out of the digital economy due to low incomes and digital literacy are able to connect to the internet and access valuable digital services, content, and e-commerce platforms for the first time. Existing internet users will also capture between a US$1.2 billion and US$4 billion consumer surplus as the result of falling broadband prices and increased perceived value of being online due to higher quality and variety of digital services and content. The full economic impact assessment is available in Annex D. 14 A SINGLE DIGITAL MARKET FOR EAST AFRICA 2.2 Drivers of growth and job Focus Box 3 Methodology and limitations creation of Economic Impact Assessment10 At its heart, the SDM initiative is about creating the economies of scale and network effects necessary for the East Africa region to be a competitive player in the global The impact assessment used panel data econometric analysis to assess the elasticities of demand for broadband access digital economy. Achieving economies of scale is the primary (that is, how sensitive demand is to price, network effects, motivation behind the region’s wider economic and political and the availability of broadband). The analysis estimated the integration agenda, creating a larger regional market to impact of integration and creation of an SDM in East Africa drive competitiveness, investment, and innovation, as well in two different scenarios: Base and High—yielding estimates as a critical mass of resources, cooperation, and attention to for the level of expected (a) network effects, (b) increase in broadband availability and adoption, and (c) reduction address shared challenges and to take advantage of shared in broadband prices. These results were in turn used to opportunities. generate estimates for expected increases in national GDP and jobs for both the Base and High scenarios. The model This is even more essential in the context of the digital used to estimate job creation leverages ITU research and the economy given the high fixed costs of telecom infrastructure impact of increased broadband penetration on job creation. and data centers, the need for tech start-ups to rapidly capture Additional econometric analysis, using panel data and a two- stage Instrumental Variable (IV) model, was conducted to market share to fend off rivals, and the exponential growth determine broadband’s impact on GDP growth, based on a in value of digital platforms and services that more users and sample divided by broadband penetration levels. This model data can generate. A deeper exploration of economies of scale also helped eliminate the risk of endogeneity and was based and network effects at each layer of the SDM framework is on a similar study conducted by the World Bank in 2017. In included in Focus Box 4. addition, consumer surplus benefits to existing users—in terms of both access to a wider network of users (network In isolation, no country in East Africa has the market size to effects) and a broader range of digital content and services— was estimated for both the Base and High scenarios. To compete with giants on the continent, such as Nigeria, much explore microeconomic impacts, detailed survey data from less global digital economy leaders such as the United States, Brazil (which was deemed comparable in terms of geographic China, and the European Union (EU). However, with a combined and population size, as well as its federal structure) was used population of nearly 180 million, an integrated East African to predict the general distributional impact of the SDM— digital market would be the ninth largest market in the world notably, how expected price decreases for mobile broadband would be distributed across lower income levels through (in terms of population), creating the market size needed to increased adoption. A brief thematic case study also explored both attract digital investment and provide a larger and more the expected increase in mobile money access, based on an familiar space for local firms to grow before having to compete interoperable market under an SDM. in the global arena. However, to realize the benefits, far more Before carrying out the analysis, various analytical models of the 180 million will have to actually be online. were explored (for example, input-output and Computable General Equilibrium (GCE) models). Ultimately, an Accelerated growth of the digital economy, both nationally econometric analysis method was adopted based on (a) the and regionally, would generate benefits for all segments of level of data availability and (b) the analytical groundwork already undertaken in previous World Bank studies. The society. For the private sector, this would open a larger market accuracy of this analysis is contingent on the quality and for entrepreneurs and small and medium enterprises (SMEs), reliability of the data used. Ideally, the models would have support the emergence of digital entrepreneurship, and provide used regionally specific data sets, but this was not available. new sources of investment. For citizens, this would increase Key data points used, sourced from the Global System for the availability and affordability of connectivity (mobile and Mobile Communications Association (GSMA), are subject to some limitations—for example, not all operators in Tanzania internet), relevant digital content and services, and create new and Burundi report figures to the GSMA. These figures are employment opportunities. Lastly, an SDM would improve the thus likely to be underestimated. quality of public services and help governments deliver these more effectively by moving them online. 10 A full account of the methodology used to conduct the economic impacts assessment, including results, is presented in Annex D. A SINGLE DIGITAL MARKET FOR EAST AFRICA 15 Figure 2: Population size of the largest global markets 1,400 1,200 POPULATION (MILLLION) 1,000 800 600 EAST AFRICA 400 200 M 200 200 200 M 0 0 E27 PAKISTAN BANGLADESH MEXICO JAPAN PHILIPPINES SOUTH SUDAN BURUNDI BRAZIL CHINA INDIA USA INDONESIA NIGERIA EAST AFRICA RUSIA TANZANIA KENYA UGANDA RWANDA Source: World Bank, 2018.11 The SDM will unlock opportunities for accelerating the wider face a risk of increased competition in regional markets by better East African integration agenda—facilitating the free movement positioned market players—early adopters—that may constrain of goods and people across borders through mutual recognition the development of their indigenous industries. However, the of digital IDs, as well as integrated digital platforms and databases economic analysis suggests that the less digitally advanced for customs, immigration, and revenue collection. The SDM will countries will gain greater access to services, infrastructure, and also help East Africa interact more effectively with the rest of investment to support the acceleration of their digital development the world, for example, by removing barriers to international and more than offset these risks. The more digitally advanced payments or the use of cloud services from other regions. nations will benefit from a much larger market for digital services and entrepreneurship, as well as lower prices and greater choice While all countries in East Africa stand to gain from an SDM, in the online services offered. While any regional integration the benefits derived from the initiative will vary depending on endeavor comes with considerations of national sovereignty, the individual countries’ level of digital maturity. Countries in the net gains far outweigh the costs and risks from joining an SDM in region currently enjoy varying levels of digital maturity (detailed terms of the development of domestic markets, job creation, and in Chapter 3). Those that are on the lower end of the spectrum the network effect of integration across the economy compared of technology adoption and have less dynamic market structures to a scenario of remaining outside of a regional integration effort. 11 Based on latest population data available from 2016. See https://data.worldbank. org/indicator/SP.POP.TOTL. 16 A SINGLE DIGITAL MARKET FOR EAST AFRICA Focus Box 4 Unlocking economies of scale and network effects Economic benefits of scale and network effects can be realized at every layer of the SDM: SINGLE CONNECTIVITY MARKET Connectivity infrastructure, such as fiber optic networks, as well as connectivity services, benefit greatly from scale, as the costs of capital investment and operating costs can be spread across more users through infrastructure sharing based on open access. Economies of scale from an SDM would enable wholesale, and subsequently, retail connectivity services prices to fall in the region as operators are able to deploy and operate regional backbone networks at lower unit cost by reaching more customers. More affordable prices would in turn help expand access and boost demand for related services, generating increased data traffic and online activity critical to the business case for further network investment in capacity upgrades and expansion to new areas. Lower cost and more accessible connectivity would also pave the way for more innovative services that rely on higher bandwidth capacity, for example, incorporating video and interactivity, further reinforcing this positive cycle. SINGLE DATA MARKET Data-based services increasingly require cloud-based data storage solutions. East Africa is currently underserved by local data centers and much of the data used and generated in the region is stored and processed overseas. A single data market would provide the scale necessary to support investment in more regionally based data centers. A single data market would enable cloud-based services to be hosted in one location yet serve the entire region, generating considerable cost savings and improving energy efficiency and security. Local/regional hosting of these services requires expensive investments in data centers as well as ancillary investments in internet exchange points (IXPs) to minimize latency and reduce reliance on international connectivity. Both would benefit significantly from the economies of scale and network effects generated by an SDM. Top-tier data centers face high initial investment costs, with low incremental expansion costs. Regional and national IXPs make more sense, the more data that is hosted locally and regionally, and where they can be co-located with larger data centers. The network effects created by a single data market would also help spur investment in data services, driving a reinforcing cycle of supply and demand for data-based services and infrastructure. SINGLE ONLINE MARKET Online services are developed in response to anticipated demand and an SDM would create a much larger potential customer base for local developers of digital services and content. Consumers in this larger market would reap corresponding benefits in terms of greater choice and service availability. Social media, e-commerce platforms, and data- driven services all rely on network effects provided by users generating content, data, and larger pools of buyers and sellers, with the value of such platforms and services growing exponentially with each new user. A critical barrier to internet adoption for many people in East Africa is the lack of content and services that are relevant to local communities and available in local languages. A larger market and compounding network effects would lead to the development of better content and could therefore also help boost uptake of digital services. ENABLING REGIONAL ENVIRONMENT Programs to promote digital skills, innovation, and investment can all benefit from increased scale and coordination across the region. An SDM would create a larger potential market for entrepreneurs while also potentially increasing access to necessary training and venture capital funds, which would be attracted by the greater growth potential afforded by a larger market. A SINGLE DIGITAL MARKET FOR EAST AFRICA 17 3. Assessment of current digital markets This section presents the background research that underpins Figure 3: Internet users and broadband the SDM Vision and Digital Roadmap proposed earlier. The penetration rates, 2017 assessment of the current state of domestic and regional digital markets, including persistent barriers to achieving an 35 SDM, presented here is also reflected in the baseline data summarized in the SDM Scorecard (found in Annex B). % OF POPULATION BASED ON SUBSCRIPTIONS 33.7 30 3.1 A single 27 26.2 26.2 connectivity market 25 21.9 Connectivity markets in East Africa are developing rapidly. 20 20 The ITU reports that there were some 33  million internet users in the region in 2017, with 8 million new internet users being added in 2016.12 Figures published by national telecom 15 regulators suggest that this figure may be even higher— 13 closer to 70 million.13 9.2 10 8.2 However, while access is growing, millions of people remain 6.7 5.2 unconnected. National markets also diverge widely in their 5 3.4 level of development, with disparities in access, pricing, 1.1 0.2 0.3 0.3 and capacity. Substantial connectivity divides persist both 0 0 0 between and within countries, with widespread inequality in BURUNDI SOUTH SUDAN TANZANIA RWANDA UGANDA KENYA access. For example, women in East Africa are 45–70 percent less likely to be internet users than men, based on country- wide surveys carried out by Facebook in Kenya, Rwanda, and Uganda.14 This presents a key challenge as well as an opportunity for regional integration. All countries in East Africa benefit from the presence of private sector telecom operators, though some domestic markets are more vibrant and competitive than others. These private sector operators have played a major role in expanding network Internet Mobile Fixed 12 Based on calculations from the ITU World Telecommunication/ICT usage rates broadband broadband Indicators database, 2017. 13 Figures from 2016. 14 Figures from 2015. Facebook, State of Connectivity 2015 Report, based on surveys conducted by Facebook over 2014–2015. See https:// fbnewsroomus.files.wordpress.com/2016/02/state-of-connectivity-2015- 2016-02-21-final.pdf. 18 A SINGLE DIGITAL MARKET FOR EAST AFRICA coverage. Nevertheless, the private sector’s investment imperative upgrades to 3G or 4G services, as the same infrastructure can is often not strong enough to support network expansion in the often be leveraged, upgrading existing regional mobile networks to region’s most remote and poor areas. In these areas, communities 3G and beyond remains a major challenge and barrier to increasing tend to be dispersed, raising the cost of deploying networks and the pool of online citizens and achieving an SDM in East Africa.1516 services. These communities also tend to have much less money to spend on connectivity services, creating limited demand for Access to internet through high-speed fiber-to-the-premises service provision. (FTTP) networks, serving business and some residential premises, remains limited and concentrated in major cities. Kenya leads Access networks, which provide connections to end users, the way in the rollout of national fiber optic broadband access predominantly rely on wireless technology in East Africa. Most networks, with four FTTP networks (Wananchi’s Zuku, Liquid current users therefore access the internet through mobile Telecom, Telkom Kenya, and AccessKenya) competing in Nairobi networks and devices. Basic 2G mobile network coverage ranges and other large cities.17 from 53 percent in South Sudan to nearly 100 percent in Rwanda. 3G network coverage ranges from 20 percent in South Sudan to 94 Innovative network and business models have been spearheaded percent in Rwanda, and 4G coverage remains limited15 (see Figure 4). to broaden access to affordable high-speed internet.18 One example includes the 2015 launch of a 120-point Wi-Fi hotspot Figure 4: 2G, 3G, and 4G mobile coverage, based network in Kampala—a partnership between Roke Telecom and on population, 2017 Google.19 However, writ large, access network availability remains constrained for rural and remote communities. 120 Currently, the expansion of access networks is primarily a national 2G 3G 4G issue being addressed through mechanisms such as national 100 broadband plans, universal service funds, and license obligations. All East African governments, except South Sudan, currently have some form of national broadband policy or strategy in place. % COVER BY POPULATION 80 Operators require a national license to build infrastructure and offer services. Similarly, wireless service providers require 60 spectrum licenses, which are issued on a national basis (although with some coordination from EACO). Greater coordination among regional players and harmonization of policy, licensing, and spectrum 40 15 See https://www.gsmaintelligence.com/markets/ - Please note that 2017 2G data point for Burundi, South Sudan, and Uganda were not published. Data present here thus reflects 2016 ITU figures for these countries. Moreover, not all operators report to GSMA. 20 16 See https://www.gsmaintelligence.com/markets/ - Please note that 2017 2G data point for Burundi, South Sudan, and Uganda were not published. Data present here thus reflects 2016 ITU figures for these countries. Moreover, not all operators report to GSMA. 0 17 See http://www.techweez.com/2016/10/28/safaricom-pilots-fiber-home-nairobi- aims-zukus-jugular/. SOUTH SUDAN BURUNDI TANZANIA UGANDA KENYA RWANDA 18 Google, for example, is experimenting with providing connectivity from a network of balloons through ‘Project Loon’; Facebook is developing solar-powered drones to relay internet signals over long distances and investing in satellite broadband technology; and Microsoft is trialling innovative uses of ‘TV white space’ spectrum to provide connectivity in Kenya and Tanzania. One ISP partner in Kenya, Mawingu, Source: GSMA Intelligence, 2017.16 is using technology to provide internet access for as little as US$3 per month. See: http://www.hope-mag.com/index.php?com=news&option=read&ca=1&a=3054. However, based on these figures, some 37 million people are 19 Facebook’s Express Wi-Fi proposition was launched in Kenya in March 2017 with 100 hotspots in Greater Nairobi and further coverage in Mombasa and Kisumu, still out of range of these mobile broadband access networks including parts of Tanzania. Partnering with Surf and Internet Solutions Kenya, at the regional level. Satellite broadband coverage is universal; Facebook provides a software and analytics stack enabling local entrepreneurs to provide access to fast and low-cost prepaid internet packages. See: https://info. however, the prices charged for related services are out of range internet.org/en/blog/2017/03/29/announcing-the-launch-of-express-wi-fi-by- for most consumers. While the presence of 2G networks facilitates facebook-in-kenya/. A SINGLE DIGITAL MARKET FOR EAST AFRICA 19 allocation in the region could help expand access networks. This thus generate significant cost savings, which could be passed on would, for example, make it easier for companies trialing new to end users, making services much more affordable.21 technologies and business models to roll them out across the region. The private sector is also spearheading the expansion of cross- Access networks are dependent on backbone networks to both border fiber links. Liquid Telecom has, for example, deployed distribute high-capacity bandwidth across each country, as well an East Africa Fiber Ring that connects five countries in the as channel international connectivity received through regional region and reaches the South Sudanese border.22 While these and global networks. Existing backbone networks primarily developments are positive, they remain isolated and fragmented, connect major cities in Kenya, Tanzania, Uganda, and Rwanda. and proprietary networks are not necessarily interconnected with Rural areas are thus also underserved by existing backbone competing networks. Further expansion of cross-border links network structures. A great deal of further investment is required could help reduce prices, particularly for landlocked countries to achieve high-capacity fiber optic backbone routes across East in the region. A coordinated regulatory response would also be Africa, which would also help introduce redundancy that limits the necessary to facilitate the interconnection of national networks to risk of outages and makes cloud-based services a reliable option. create seamless regional backbone networks. An interconnected network, with greater infrastructure sharing and coordination, Typically, the expansion of networks is driven at a national level would both decrease the cost of infrastructure deployment and and is private sector led, though some government support may enhance competition that could help drive down prices further. be necessary for coverage of areas where deployment would A more integrated market would also provide the scale needed otherwise be unprofitable (as noted earlier). PPPs are an effective to attract greater investment in fiber optic infrastructure, which tool that can be leveraged to help fund the development of could help increase deployment and access. backbone networks in the region, yet these partnerships must be designed in a way that maximizes impact and value-for-money. In addition to the telecom networks themselves, digital services Nevertheless, in countries where the government has retained rely on supporting infrastructure such as IXPs and data centers a monopoly of national backbone networks, such as Tanzania, that can support the local or regional exchange of internet this model would not be feasible. While governments in Burundi traffic and origination of content. This will become increasingly and Uganda have also retained ownership of some backbone important as internet usage becomes more bandwidth intensive infrastructure, these networks face some private competition. and sensitive to latency. IXPs play an important role in exchanging Notably, a lack of competition in network ownership has ripple traffic among ISPs, as well as between ISPs and content providers, effects on the wholesale market prices for broadband (more on so that local traffic does not have to be exchanged abroad using this in the following paragraphs). expensive international connections. Meanwhile, data centers can be used to cache international content, to host local traffic, The vast majority of capacity connecting the region to the rest and as nodes of content delivery networks (CDNs) and even IXPs. of the world is currently being supplied through submarine Data center infrastructure thus also plays an important role in cables, landing in Kenya and Tanzania, supplemented by small decreasing the cost and latency of accessing content. amounts of international capacity provided through cross-border terrestrial cables, as well as some satellite broadband capacity. There is currently a lack of supporting infrastructure in most of In other words, only two of the six countries in the region benefit East Africa. A mere six major commercial data centers are located from direct access to international capacity through submarine in East Africa (five in Kenya and one in Tanzania), out of some cables. The other four landlocked countries must access 4,124 co-locations reported globally.23 The largest is the East international capacity through cross-border terrestrial cables, Africa Data Center, a carrier-neutral facility, built by a subsidiary of with corresponding markups in pricing. This partly helps explain Liquid Telecom. While the demand for regional data centers and disparities between existing national connectivity markets. Up to 21 Users in landlocked countries in Africa pay on average US$232 more per month 90 percent of all internet traffic in East African countries is currently for fixed broadband access than those living in coastal areas, according to the international.20 Reducing the cost of international transit would World Bank’s 2016 Digital Dividends report. See http://www.worldbank.org/en/ publication/wdr2016. 20 According to the stakeholder consultations conducted. 22 See https://www.liquidtelecom.com/about-us/network-map.html. 23 We note that this is not a comprehensive list. See http://www.datacentermap. com/datacenters.html. 20 A SINGLE DIGITAL MARKET FOR EAST AFRICA IXPs is currently limited, partly due to the high cost of connectivity, border network deployments. The de facto monopoly of the the scope and usage of this supporting infrastructure is expected government-owned national ICT backbone in Tanzania, including to grow in tandem with increased demand for content. This the lack of adequate redundancy, and service reliability of the technology would, for example, be a prerequisite for the backbone infrastructure, have a particularly pernicious impact development of more advanced digital services. on wholesale connectivity competition, pricing, and quality for landlocked countries in the region, which rely on networks in Most backbone networks in East Africa offer wholesale either Kenya or Tanzania for access to submarine cables. connectivity to telecom operators, ISPs, and government and business customers. Currently, wholesale costs are amplified by All countries in East Africa have at least two sizable retail the high charges associated with international IP transit services, operators. However, the retail markets in Burundi and Kenya especially in landlocked countries, required to facilitate cross- are among the least competitive, with a single operator holding border transfers of connectivity from costal landing points (as a majority market share (see Figure 5). In contrast to its state- noted earlier). At present, use of international IP transit services owned monopoly wholesale backbone market, Tanzania’s retail generally tends to be high in the region, as most of the content market is highly competitive, with its top three retail operators being accessed currently originates from outside of the region. enjoying a roughly equal market share and three smaller mobile This would, however, change with the development of more operators fiercely cutting prices to gain a foothold. Increased local content. competition in the retail market in the other East African countries could similarly help drive down prices for consumers and also Moreover, if wholesale customers wish to connect to a data incentivize innovation of new services such as mobile money. center or peer with an IXP in another country, they currently face higher charges, because they need to purchase capacity Figure 5: Market share of mobile market from two separate national backbone networks. Current connections among retail operators licensing arrangements associated with purchasing capacity from backbone networks therefore contribute to limited and high-cost 100% cross-border connectivity. 90% A single connectivity market would aim to minimize any cost 80% differentials in connecting to infrastructure within or across 70% countries in the region. It would also seek to harmonize licensing provisions, making it easier to access wholesale capacity across 60% borders, which would enable new regional wholesale players 50% to emerge that would spur increased competition. Limited competition in the wholesale market currently contributes 40% to high wholesale prices. Other factors presently affecting 30% wholesale prices include the high cost of power, operations and maintenance, taxes on both revenues and rights-of-way,24 and 20% cost of licensing fees. 10% High wholesale prices directly trickle down to retail prices, 0 adversely affecting the affordability of internet services for BURUNDI RWANDA SOUTH SUDAN TANZANIA UGANDA KENYA consumers. Bringing down these costs would thus have a positive impact on the retail market. Beyond pricing, the quality and reliability of connectivity can be an issue across East Africa. Service outages are not uncommon in the region, which partly stem from Operator 1 Operator 2 Operator 3 Operator 4 a lack of coordination with respect to infrastructure and cross- Operator 5 Operator 6 Operator 7 Operator 8 24 Rights-of-way charges for deploying fiber can act as a barrier to infrastructure deployment and result in higher prices for services offered over fiber networks. In Tanzania, for example, the Tanzania National Roads Agency quotes rights-of- Source: GSMA Intelligence, 2017. way charges of US$1,000 per kilometre per year. A SINGLE DIGITAL MARKET FOR EAST AFRICA 21 While retail prices have improved significantly in recent years, However, as noted earlier, this will take some time where network they remain unaffordable for millions in the region, especially coverage is limited. Moreover, it will require widespread uptake of low-income households at the base of the pyramid. A 500 MB 4G-capable handsets, whereas a majority of users currently own mobile broadband package currently costs between 3 percent only basic phones. and 22.3 percent of gross national income (GNI) per capita. The highest costs are found in Burundi, followed by South Sudan, and The affordability of devices, particularly internet-compatible the lowest is Tanzania and Kenya (see Figure 6). smartphones, is a major barrier to connectivity in the region. Devices have become more affordable in the last five years, due Figure 6: Mobile broadband prices as % of to the introduction of low-cost Chinese brands. Jumia in Uganda GNI per capita reports that the average cost of smartphones has fallen by 45 percent, between 2014 and 2017, reaching US$99,25 and appears to have fallen as low as US$40 (approximately) in some markets. 50 Nevertheless, taxes on mobile devices, such as VAT and customs duties remain high in many parts of the region. For example, 500MB 1G taxes represented 18 percent to 19 percent of the retail price 40 39.2 in Tanzania and Uganda in 2016.26 A brief period of zero VAT on handsets in Kenya in 2009 saw the sales of devices and mobile services soar. This also had a positive net impact on government revenues, due to higher levels of telecom service tax collected and 30 positive impacts on the overall economic growth rate. However, 27.9 % OF GNI PC 26.1 this measure adversely affected sales in neighboring markets and triggered cross-border smuggling. Lower, yet harmonized, 22.3 tax policy could thus prove effective in expanding access to 20 17.5 affordable devices. 14.1 In a single connectivity market, it should also be affordable for 10 citizens to use connectivity services when travelling within the 10 region. As noted earlier, the ONA has eliminated roaming charges 4.4 4.4 4.4 5.4 for voice calls and SMS messages between several EAC countries. 3 Despite the success of this initiative, plans to extend it to data, 0 SMS, and mobile money services have been slow to materialize. Kenya, Uganda, and Rwanda have now adopted the ONA for data BURUNDI RWANDA SOUTH SUDAN TANZANIA UGANDA KENYA but with nonuniversal application. Tanzania and Burundi have made commitments to join the ONA for voice but have not yet followed through as of the time of publication. Lowering regional retail prices will be a critical first step toward Source: ITU, 2017. increasing demand for connectivity, creating the user-base necessary for the development of a single data and online service Retail pricing in a competitive environment is predominantly market. Demand for connectivity will also rely on there being driven by the underlying costs faced by retail service providers, more attractive online content and services available, as well as including licensing fees, network costs, wholesale prices, and prospective users having the digital skills to access them. These taxes. Retail services in a single connectivity market would benefit issues are discussed in subsequent sections of this chapter. from the improvements in regional infrastructure and wholesale 25 See http://www.techjaja.com/jumias-2017-white-paper-gives-us-details-mobile- service market. Equally, network costs should decrease where internet-penetration-trends-ugandan-online-consumer/. upgrades to higher capacity networks, such as 4G, are made. 26 See https://www.gsmaintelligence.com/research/2017/07/taxing-mobile- connectivity-in-sub-saharan-africa/630/. 22 A SINGLE DIGITAL MARKET FOR EAST AFRICA 3.2 A single data market “ The ability to easily share data across borders, be it for customs, Data sharing can be immigration, or e-commerce purposes, will be a key prerequisite for developing an SDM. For example, when making an online facilitated through both e-commerce transaction, companies may need to query national digital IDs. Data sharing can be facilitated through both the the digitization and digitization and standardization of data. However, for data to flow freely, there can be no undue restrictions on the storage and standardization of data. processing of data outside national borders. The NCIP has a working group on intergovernmental data sharing. Through this initiative, Kenya, Rwanda, and Uganda have reportedly and Uganda.30 Hence, some content hosted in Kenya may not agreed to Harmonize the related regulatory framework, enabling be legal in other countries in the region. Recently announced data sharing on common terms. South Sudan is in the process of regulations in Tanzania require a US$930 license fee for bloggers, developing similar regulation. This framework defines both which putting the practice out of reach for most citizens if they operate data can be shared and what protocols must be followed. It will, in compliance with the law.31 For a regional content ecosystem for example, allow for the integration of national ID and sim card to thrive, it is important to ensure legal clarity for those registration databases. developing, transmitting, or hosting local content and to enable rather than suppress such content development. This is equally Data localization laws, requiring data on local citizens to be important in terms of encouraging investment from multinational processed and stored locally, can be detrimental to data sharing, content companies, such as Google. Companies wishing to host and thus have a crippling impact on innovation. Moreover, it third-party content in the region will also require protection can increase the costs of data storage, as well as limit choice against intermediary liability. For example, carrier-neutral data for users and reduce services available.27 Currently, no country centers should be able to host content without being held liable in East Africa has a data localization law, however, some sector- for how third parties use their services.32 specific regulations are enforcing data localization. For example, MTN in Rwanda was recently fined US$8.5 million (10 percent Moreover, to promote the free flow of data across the region, of its annual turnover) for maintaining Rwandan customers’ companies need to be assured that intellectual property is in Uganda.28 However, Rwanda has also introduced a law that protected, both domestically and across borders through the obliges government departments to maintain a cache of their development of a regional intellectual property environment. data within national borders, which is helping reduce costs for This will also encourage more businesses to offer services international data transport. Overall, facilitating data portability across the region. Initially, this would mean that intellectual will be an important element of the SDM, which also helps stem property registered in one country would be recognized data Monopolization / Monopolies.29 and enforced across the region. In the longer term, a single regional registration process could be defined and aligned Similarly, data content restrictions can prevent data from flowing with international agreements relating to the registration and freely across borders and present a key constraint to an SDM. enforcement of patents, trademarks, and copyright. A regional Most countries appear to have similar approaches to content body that manages registration and enforcement could also help restrictions and filtering. However, a 2016 ‘Freedom on the Net’ facilitate this. survey, which covered Uganda, Rwanda, and Kenya, pointed to the internet being ‘free’ in Kenya but only ‘partly free’ in Rwanda 30 Freedom House. 2016.Freedom on the Net Report, at https://freedomhouse. org/report/freedom-net/freedom-net-2016. While the other countries were not 27 See https://www2.deloitte.com/content/dam/Deloitte/sg/Documents/about- covered in this survey, the Freedom House Freedom of the Press report states deloitte/sea-about-aec-digital-economy-free-flow-of-data-2016.pdf . A 2016 that the press is either ‘partly free’ or ‘not free’. These restrictions may extend report by Deloitte outlines the potentially damaging effect of data localization. to online content as well. See https://freedomhouse.org/report/freedom-press/ freedom-press-2017. 28 See https://www.cnbcafrica.com/news/2017/05/17/rwanda-utilities-regulatory- authority-fines-mtn-us-85m-non-compliance/. 31 See https://www.ft.com/content/36098722-4623-11e8-8ae9-4b5ddcca99b3. 29 See Single Digital Market for Africa Report. Transform Africa Summit May 2018. 32 See https://www.internetsociety.org/doc/promoting-content-africa. A SINGLE DIGITAL MARKET FOR EAST AFRICA 23 Figure 7: Date of entry into force for a selection of agreements and treaties regulating the international protection of intellectual property Year entering Paris Berne Patent Madrid WIPO into force Convention Convention Cooperation Protocol Copyright Treaty Treaty Burundi 1977 2016 û û 2016 Kenya 1965 1993 1994 1998 û (signed only) Rwanda 1984 1984 2011 2013 û South Sudan û û û û û Uganda 1965 û 1995 û û Tanzania 1963 1994 1999 û û Number of 177 176 152 101 96 contracting countries Subject matter International International Simultaneous filing International Copyright protection protection of protection of of international registration of in the digital patent and copyright patents trademarks environment trademark rights Source: World Intellectual Property Organization, 2018.33 Note: WIPO = World Intellectual Property Organization. Countries in Africa face challenges relating to registering purchases.35 Harmonized data protection and privacy procedures intellectual property, as the registration process tends to be can also avoid a “race to the bottom,” in terms of standards, in complex. Enforcement of intellectual property protection rules a bid to attract data-driven services at the expense of consumer when infringements occur is also not always effective. Compared protection.36 to other countries in Africa, Kenya appears to have a fairly Data protection and privacy laws are designed to regulate advanced intellectual property environment, with dedicated the collection, transmission, storage, and usage of personal bodies, frameworks, and enforcement agencies to manage cases data. However, only 40 percent of countries on the continent of intellectual property infringement, and a new online copyright have implemented comprehensive laws on the issue—none in registration system. However, many weaknesses remain, including East Africa. Bills have been proposed in both Kenya (2013) and gaps in digital copyright protection. Enforcement bodies also tend Uganda (2015); however, they diverge widely and are considered to lack resources, which limits their efficacy.34 inadequate in terms of supporting personal data storage and cross- While data needs to be stored and processed across borders, it also border data transfer.37 There is thus scope to improve national needs to be done safely. A common approach to data protection 35 2016 Centre for International Governance Innovation (CIGI) and Ipsos Global Survey on Internet Security and Trust, at https://www.cigionline.org/internet- and privacy is needed for there to be trust in online services survey-2016 Global Survey on Internet Security and Trust. The CIGI-Ipsos survey and sufficient safeguards in place to protect all consumers. polled over 24,000 users in 24 countries about their attitudes on privacy and security. When asked about how concerns regarding online privacy have changed Survey data from Kenya suggests that privacy concerns may be their online behavior, 47 percent of respondents in Kenya reported that they deterring the public from making online financial transactions and made fewer financial transactions online than they used to and 44 percent were making fewer online purchases. 33 See http://www.wipo.int/treaties/en/summary.jsp. 36 See Single Digital Market for Africa Report. Transform Africa Summit May 2018. 34 See http://www.theglobalipcenter.com/wp-content/uploads/2017/02/GIPC_IP_ 37 See http://unctad.org/en/Pages/DTL/STI_and_ICTs/ICT4D-Legislation/eCom- Index_2017_Report.pdf. Data-Protection-Laws.aspx. 24 A SINGLE DIGITAL MARKET FOR EAST AFRICA policy in this area and develop a coordinated regional approach any regional roadblocks to related initiatives. The role and that is grounded in best practice and adopts common standards. availability of open data at a regional level could be compounded by standardizing approaches to data collection, formatting, The EAC introduced a legal framework for cybersecurity, published and publication through harmonized data classification policy. in 2008, which includes provision on privacy and data protection. Regional best practice is currently being supported by the African However, these legal provisions have not been fully adopted. A Development Bank. single EAC law on data protection, based on the Convention on Cybercrime and Personal Data Protection of the African Union Meanwhile, the threat of cybercrime in the region is growing (AU) (adopted in 2014), could be an effective means of improving rapidly. Both governments and businesses have reported related and harmonizing existing regulation.38 Guidelines were recently incidents and data breaches.44 The full scale of the problem published to aid effective implementation. As of July 2018, eleven is unknown, partly due to inadequate detection capabilities. countries have signed and three have ratified the AU Convention While digitization offers promising opportunities, it can also though none from East Africa.39 Related provision ought to be create vulnerability. This phenomenon also poses challenges swiftly ratified by countries in the region. It may also be prudent toward regional integration as some governments may perceive to align regional regulation with standards applied in other major an SDM to result in less national-level control and protection in markets, to facilitate global data exchange. Examples include the cybersecurity. Yet it also presents an opportunity to maximize EU‘s General Data Protection Regulation (GDPR) and Network and resources and capacity in fighting cybercrime. Information Security Directive (NISD), introduced in 2018, which precludes data exchange outside the EU with counterparts that Cybersecurity laws have been passed in Kenya, Rwanda, South do not meet comparable data protection standards.40 Sudan, and Uganda. Draft legislation also exists in the other EAC countries. Most are based on the 2008 EAC framework, Consideration must be given to data protection in the public noted earlier, which seeks to harmonize cyber laws regionally. sector, as well as the private sector. Government agencies are Member states of the NCIP have also signed a memorandum increasingly collecting and digitally storing large amounts of of understanding on cybersecurity, dating from 2014, which citizen data through initiatives such as national ID schemes. This provides a collaborative platform for jointly tackling cybercrime. data needs to be properly protected and shielded from abuse. It calls for the establishment of a regional incident response team Rwanda’s data collection policy has, for example, been criticized and includes provisions on information sharing and collectively for not adequality protecting citizens’ data.41 developing the technical capabilities needed to detect, prevent, and respond to cybersecurity threats.45 Conversely, increased availability of open data, that anyone can access, use, and share, is helping spur service delivery Cyber laws need to be regularly reviewed and updated, given the improvements and increase the transparency of public agencies. constantly evolving nature of threats. It is in all countries’ interests Open data is also being leveraged to provide data-driven insight to support a coordinated regional response to cybersecurity, that supports innovation across a wide range of sectors, which is building on these provisions and supporting neighboring in turn helping address various socioeconomic issues—examples countries where required. Pooling resources is not only cost- include enterprises such as aWhere42 and Medafrica,43 which are effective, as in an integrated data market, any one country is only using open data to help smallholder farmers increase their yields or as protected as its weakest link. As more and more digital services improve access to health information. Open data initiatives should evolve, the threat will also intensify, with potentially detrimental therefore be encouraged, based on best practice, addressing effects on regional e-commerce and e-health services or critical 38 See http://unctad.org/en/PublicationsLibrary/dtlstict2016d1_en.pdf. infrastructure such as power grids, financial and customs systems, 39 See https://au.int/sites/default/files/treaties/29560-sl-african_union_ and digital ID schemes that rely on data and digital services. convention_on_cyber_security_and_personal_data_protection_1.pdf. 40 See https://eugdpr.org/the-regulation/. 44 See https://www.iafrikan.com/2016/10/19/a-burundian-hacker-was-able-to-get- 41 See https://privacyinternational.org/advocacy-briefing/771/right-privacy-rwanda. the-details-of-over-500-000-kcb-kenya-customers-through-a-vulnerability-in-the- 42 See www.awhere.com. banks-app/. 43 See www.shimbamobile.com/. 45 See https://www.nciprojects.org/project/ict-infrastructure. A SINGLE DIGITAL MARKET FOR EAST AFRICA 25 3.3 Single online market access to capital are able to bear these costs and launch across multiple countries and grow their market share, whereas smaller Online services cover a wide range of online activities, including regional start-ups may not have the access to capital, time, or digital education and health services, access to information, news expertise to do so, putting them at a distinct disadvantage and and entertainment, online shopping, and social networking. blunting competition. The online service ecosystem in East Africa is growing quickly, Digital identification is an important enabler of many online bringing wide-ranging social and economic benefits for citizens services that require providers to know their customer, and businesses. Many new local e-commerce companies are significantly reducing the risk of fraud and increasing trust and emerging, such as Ugandan Dondolo, 2fumbe, and Masikini46 or efficiency on both sides of the transactions. Digital IDs can, for Rwandan Yubeyi Online Shopping and GroceWheels.47 Roughly example, support trade; digital payment platforms; e-government 7 percent of Kenyans were using e-commerce services in 2015, services; financial inclusion; and access to health, education, and generating approximately US$47 million in revenue.48 While, Kenya social protection programs. Yet, an estimated 71 million people in leads the way in e-commerce, related transactions only accounted the EAC are unregistered, corresponding to 37 percent of the total for roughly 5 percent of all transactions made, suggesting that regional population, preventing them from obtaining an official there is much more room for growth in this sector.49 means of identification. However, the number of unregistered A majority of e-services currently offered in East Africa are not people improves somewhat over the mandated age-limit for home grown and still originate outside of the region. Notably, enrollment in voting or national ID schemes—excluding 16.1 most popular e-commerce websites are also registered with an million people in the region.52 In Rwanda, everyone over 16 years international, not a local, domain name.50 A number of websites is said to be registered, whereas in South Sudan 43.7 percent of also operate across multiple countries in East Africa, but with local the population over the age of 18 is unregistered.53 domains registered in each country, suggesting that these sites Figure 8: Share of population that is are tailored to each country rather than a wider regional market. unregistered Although some online services operate across East Africa, they 60 face many cross-border trade barriers, which are reflected in % OF PEOPLE THAT ARE UNREGISTERED both their front-end (for example, websites) and back-end (for 53 Total Population 49 47 example, warehousing and logistics) operations. This is a key Population binding constraint to an emergence of an SDM. 44 above cut-o 40 A good example of this is Jumia, a pan-African online e-commerce 30 platform, selling local and international goods, founded in 27 Nigeria. Jumia now operates in some 23 countries, including 24 Kenya, Burundi, Rwanda, Tanzania, and Uganda. Despite its 19 20 18 20 regional footprint, Jumia operates separate websites, warehouse systems, and delivery teams in each country, and each website only offers domestic shipping.51 Large foreign firms with easy 4 0 0 46 See https://www.export.gov/article?id=Uganda-eCommerce. SOUTH SUDAN BURUNDI UGANDA TANZANIA RWANDA KENYA 47 See http://www.newtimes.co.rw/section/read/228485. 48 See https://www.oxfordbusinessgroup.com/overview/bright-outlook-improving- economic-conditions-and-changing-consumer-preferences-point-strong-years. 49 See http://www.monitor.co.ug/Business/Prosper/Uganda-advised-consider-e- commerce/688616-4329744-tww6d0/index.html. Source: World Bank Identification for Development (ID4D) data set, 2018. 50 Ranking based on visits over a three-month period from visitors with the Alexa toolbar installed. While it is true that a local company could also use .com, the national domain is typically used for sites serving the national market. 51 See http://www.newyorker.com/business/currency/e-commerces-african- 52 Referred to in the ID4D data set as the ‘cut off-age’: age 16 in Rwanda and challenge-selling-to-people-who-arent-online; http://jumia.pr.co/99764-africa-s- Burundi and age 18 in South Sudan, Uganda, Tanzania, and Kenya. leading-e-commerce-site-jumia-has-now-launched-in-senegal. 53 Calculation based on the World Ban ID4D data set, 2018—please note that the year data was collected varies. 26 A SINGLE DIGITAL MARKET FOR EAST AFRICA To date, all six countries (at least partially) use digital IDs or are Figure 9: Penetration of financial institution in the process of developing them. These digital ID schemes accounts versus mobile money accounts typically include biometric data such as fingerprints and iris scans, used to prevent fraud and duplication, and relevant data is stored 80 73 Financial institution account digitally on a national database within each country. Where digital ID initiatives are being deployed in East Africa, they remain 56 Mobile money account % OF ADULTS OVER 60 primarily focused on national-level solutions, with at least some THE AGE OF 15 51 corresponding barriers to cross-border use. 40 37 39 East Africa is the only region in Africa that currently allows for any 31 33 level of reciprocal recognition of IDs, albeit for physical rather 21 20 than digital IDs, in relation to travel within the EAC. The World Bank’s ID4D initiative has an ongoing program in East Africa to help 9 7 0 1 accelerate regional integration of IDs and associated databases, supporting mutual recognition and further interoperability.54 BURUNDI RWANDA SOUTH SUDAN TANZANIA UGANDA KENYA Greater integration of other databases, which leverage national ID numbers, could be used to check consumers’ credit ratings, improving the provision of financial services across the region. A harmonized regional approach to IDs would thus also help support Source: Findex, 2017. the development of a regional digital payments ecosystem. Note: The latest Findex survey data from Burundi was collected in 2014. Mobile Until recently, East Africa has predominantly been a cash money adoption in Burundi is expected to be much higher than figures indicated here as it is known to have grown in recent years. economy. The region has seen low penetration of credit cards and bank accounts, associated with traditional banking. The While mobile money is popular, few online services provide direct mobile money revolution, initially spurred by M-PESA, but now online mobile money payment capability. Nevertheless, digital offered by many providers across the region, has changed this payments are growing quickly in popularity within some markets— dramatically and is enabling a new wave of financial inclusion usage of digital payments among adults to send or receive money that has now spread beyond the region and globally.55 Mobile in Uganda is 55 percent and 79 percent in Kenya but remains money is also seen as having a positive impact in helping close limited in South Sudan and Burundi.59 Moreover, there are several the gender gap in both access to finance56 and supporting access barriers to making cross-border payments, including network to economic opportunities.57 Most East African countries have interoperability and high transaction costs, which in turn restrict significantly more mobile money accounts users than account the development of regional online services and the growth of holders with other financial institutions. The exceptions are South e-commerce (for which they are a prerequisite). Sudan, where mobile money is not yet licensed, and Rwanda.58 Mobile payment systems in East Africa are mainly domestic, limited to a single operator’s network. These systems typically have limited interoperability with other domestic operators, let alone other regional and international operators.60 Kenya’s Safaricom is 54 World Bank ID4D documentation, 2017. one of the few operators to have enabled international money 55 See The Global Findex Database 2017 Measuring Financial Inclusion and the Fintech Revolution. transfers. Registered M-PESA subscribers can send money to 56 There is no large gender gap in access to mobile money in Kenya and Uganda. Vodacom Tanzania, MTN Uganda, and MTN Rwanda subscribers. See Findex data 2017. Transfers are received in local currency, with conversion rates and 57 A study in Kenya found that access to mobile money services enabled women- headed households to increase their savings by over 20 percent; allowed 185,000 fees advertised in a text message before money is transferred. women to develop business or retail activities, reducing extreme poverty among women-headed households by 22 percent. See: Suri, Tavneet, and William Jack. 2016. “The Long-Run Poverty and Gender Impacts of Mobile Money.” Science 354 (6317): 1288–92. 59 See The Global Findex Database 2017 Measuring Financial Inclusion and the Fintech Revolution. 58 See https://www.gsma.com/mobilefordevelopment/programme/mobile-money/ global-adoption-mobile-money-2015-look-data#_ftnref1. 60 International e-commerce in Africa: The Way Forward, International Trade Centre, 2015. A SINGLE DIGITAL MARKET FOR EAST AFRICA 27 Figure 10: Mobile payment systems deployed in the EAC System Burundi Kenya Rwanda South Sudan Tanzania Uganda Africell Money ü Airtel Money ü ü ü BK mVISA ü Dau-Pesa ü EcoKash ü EzeeMoney ü ezyPesa ü Leo Manoti ü M-PESA ü mHose ü M-Sente ü mCash ü Micropay ü MobiCash ü ü ü MTN Mob’ Money ü ü Orange Money ü PAYG Platform ü Tangaza Pesa ü Tigo Cash ü Tigo Pesa ü Vodacom M-pesa ü Source: GSMA Mobile Money Deployment Tracker, 2017. 61 In other words, no interoperable mobile payment system has first and foremost be a technical one, policy makers can also yet been deployed across the entirety of East Africa. Equally support the process by mandating interoperability. The NCIP’s important, the transaction fees for doing so remain very high ONA roaming initiative offers a model and platform for doing so. (both between domestic platforms and across borders between Related proposals to eliminate roaming charges for mobile money platforms), suppressing transactions even where technically could also include targeted tax reductions. Kenya, Tanzania, and available. Greater interoperability as well as lower cross-platform Uganda currently all apply a 10 percent tax on mobile money and cross-border transaction fees could foster a regional transaction fees.63 digital payment ecosystem. The World Bank is helping develop supporting ICT infrastructure that links regional stock exchanges Higher transaction fees for cross-border payments can also across the EAC, as well as settlement and depository facilities, be a deterrent to making international transfers. For M-PESA which could facilitate interoperability.62 While this process will customers to send K Sh 1,000 (~US$10) within Kenya, they incur a cost of K Sh 15 (1.5 percent), whereas the equivalent international 61 Data excludes those firms that are not GSMA members (for example, Lumicash in Burundi, Halotel in Tanzania, and so on). 62 See ‘Financial Sector Development and Regionalization Project I’, World Bank, 63 See https://www.gsma.com/mobilefordevelopment/wp-content/ September 2016. uploads/2016/07/Digital-Inclusion-and-Mobile-Sector-Taxation-2016.pdf. 28 A SINGLE DIGITAL MARKET FOR EAST AFRICA transaction costs K Sh 33 (3.3 percent).64 The cost of transaction is their incentive to develop local content.68 Regional payment also affected by the high currency exchange rates applied in East platforms thus need to be interconnected with global platforms Africa, which for example, inhibit small e-commerce transactions such as PayPal. A large single market can create the business where the extra costs can appear prohibitive. incentive needed for global payments platforms to facilitate greater international financial inclusion. Central banks in East Africa have not historically operated reciprocal accounts, which means that most transactions must A prerequisite for conducting commercial transactions online be exchanged from the local currency into U.S. dollars and then is the presence of e-transaction laws that recognize the legal into the new local currency. Payments thus incur two sets of equivalence between paper-based and electronic forms of conversion charges. Steps are being taken to facilitate international exchange. Equivalence laws have been passed in Kenya, Rwanda, transfers. The Regional Payment and Settlement System (REPSS) Tanzania, and Uganda and draft legislation exists in Burundi. The was introduced in 2012, with the aim of facilitating payments laws have been developed under the EAC’s Legal Framework for between 19 member states in the Common Market for Eastern Cyberlaws.69 In October 2015, the East African Legislative Assembly and Southern Africa (COMESA) region. This initiative includes also passed an EAC Electronic Transactions Bill, which sought to Uganda, Rwanda, and Kenya. Burundi also aims to implement the develop a safe, secure, and effective environment for consumers, initiative. The REPSS provides a single gateway for the settlement businesses, and governments to conduct and use electronic of international transactions in euros or U.S. dollars, with the transactions. It sets regional standards in relation to electronic Bank of Mauritius providing reconciliation services.65 Moreover, signatures, e-government services, consumer protection, and the the East Africa Payment System (EAPS), established in 2014, limitation of liability of service providers.70 Alignment of national is enabling the central banks in Kenya, Uganda, Rwanda, and laws with the EAC bill is mixed, with some deviations in each country Tanzania to freely interchange local currencies. The EAPS links the and lack of sufficient regional harmonization/perspective in many. real-time gross settlement system (RTGS) of the central banks in the four countries and includes reciprocal accounts between the In addition to the cost and practicalities of making digital banks that allow for transfers in local currency.66 payments, consumer familiarity and trust in digital payment systems is a major barrier to greater adoption. This issue is more The EAC has also articulated plans for a monetary union and aims pronounced for cross-border payments, as the potential recourse to adopt a single currency for East Africa by 2024. These plans for fraud or failed transactions becomes more complicated also feature the harmonization of financial systems, accounting and costly. E-transactions laws can help formalize the payment and reporting practices, and the establishment of an East African process. However, consumer protection legislation relating to Central Bank.67 The World Bank is currently helping support e-commerce is an important factor in developing consumer greater interoperability of capital markets. If successful, these trust in e-commerce services. In East Africa, Kenya, Rwanda, and efforts will remove the necessity for regional currency exchange Uganda have online consumer protection laws in place, with draft and associated charges. In the meantime, current efforts to lower legislation in place in Burundi and Tanzania.71 However, these the cost of currency exchange rates between countries in the laws need to be harmonized across borders, including reciprocal region ought to be fully adopted and implemented. agreements to recognize and enforce each other’s approaches. Countries in the region are currently committed to doing so under International financial restrictions are also limiting some the EAC Electronic Transactions Bill yet incur no penalties for payments to the region, which can have ripple effects on local noncompliance.72 digital entrepreneurship. For example, Google Play allows app developers to upload apps and sell them anywhere in the 68 See https://www.internetsociety.org/resources/doc/2015/discussion-paper- the-mobile-app-divide/. The reasons for this lack of international payments is world. However, currently local developers are unable to receive relatively opaque. payments for new apps uploaded on this platform, which reduces 69 See http://unctad.org/en/Pages/DTL/STI_and_ICTs/ICT4D-Legislation/eCom- Transactions-Laws.aspx. 64 See https://www.safaricom.co.ke/faqs/faq/279. 70 See http://www.eala.org/new/index.php/media-centre/press-releases/873- region-set-to-realize-e-business-growth-as-eala-enacts-eac-electronic- 65 See https://www.centralbank.go.ke/national-payments-system/. transactions-bill. 66 Burundi and South Sudan are yet to develop RTGS systems. See http://www. 71 See http://unctad.org/en/Pages/DTL/STI_and_ICTs/ICT4D-Legislation/eCom- theeastafrican.co.ke/news/EAC--is-a-single-currency-on-the-way-/2558-3323198- Consumer-Protection-Laws.aspx. view-printVersion-89y0bbz/index.html. 72 See http://www.cofek.co.ke/EAC%20Electronic%20Transactions%20Bill%202014. 67 See http://www.eac.int/integration-pillars/monetary-union. pdf. A SINGLE DIGITAL MARKET FOR EAST AFRICA 29 People in East Africa are used to face-to-face transactions. However, businesses also have a significant role to play in terms However, boosting consumer trust in e-payments and virtual of boosting consumer confidence in the safety of online services, transactions will be critical to supporting the development of a by integrating service functions that build trust as part of their single online market. This may, for example, require government business model. This means enabling easy and transparent initiatives that encourage citizens to make digital transactions for returns, providing customer ratings and review systems and the first time, through the introduction of digital public services. escrow accounts so that payments are not taken until goods Digitizing both internal and external functions can also increase are received. An SDM would allow e-commerce enterprises that government efficiency and improve the quality and reach of deliver good customer services to expand and disseminate best public services. As noted earlier, the digitization of government practices across the region. functions and data can also facilitate data sharing, with the adoption of regional standards for ICT hardware and software For e-commerce to thrive in the region, goods and services will that ensure interoperability wherever possible. Moreover, it will need to be delivered seamlessly across the region, at scale. This make it easier for citizens and businesses to interact with relevant will require the gradual harmonization and streamlining of regional public services across the region, by making public authorities trade and customs policies, facilitating cross-border shipments of more accessible online and increasing the reach of both public physical goods. The EAC has already adopted a Common Market services as well as online services more broadly. Protocol (CMP) that came into force in 2010, which promotes the free movement of goods within the region by eliminating tariff Most countries in East Africa still have a long way to go in terms and nontariff barriers.74 The Community Customs Management of developing their e-government capabilities (see Figure 11). (Amendment) Bill75 and the EAC Elimination of Non-Tariff Barriers However, regional coordination can accelerate the proliferation of Bill,76 both introduced in 2015, also support this objective. digital public services through the sharing of ideas and platforms. Figure 12: Good country scores Figure 11: E-Government Development Index scores (measured on scale from 0 to 1) 100 97 2014 90 85 93 91 91 90 0.8 79 82 2016 80 0.72 70 69 69 60 0.6 50 40 30 0.42 20 0.4 0.36 0.34 10 0.34 0.29 0 0.23 BURUNDI RWANDA UGANDA TANZANIA KENYA 0.2 0.18 0 Source: Common Market Scorecard, 2016.77 BURUNDI SOUTH SUDAN EUROPE KENYA RWANDA TANZANIA UGANDA AFRICA 74 See https://www.eac.int/common-market. 75 See http://www.eala.org/documents/view/the-east-african-community-customs- managment-amendment-bill2015. 76 Ibid. Source: United Nations 2016. 73 77 Scored 0 to 100, with 100 being the best. These are composite scores, depicting performance against customs and trade liberalization provisions agreed under 73 The E-Government Development Index is calculated based on the proportion the EAC CMP. In other words, they reflect the removal of trade and customs of central government employees and organizations making regular use of barriers, facilitating cross-border movement of goods. Latest scores were computers, the internet, local area networks, and intranets and making their compiled in 2016. services available to citizens online. Latest scores available are from 2016. See See https://d3n8a8pro7vhmx.cloudfront.net/eatradehub/pages/2893/ https://publicadministration.un.org/egovkb/en-us/reports/un-e-government- attachments/original/1481012380/East-Africa-Common-Market-Scorecard-2016. survey-2016. pdf?1481012380. 30 A SINGLE DIGITAL MARKET FOR EAST AFRICA However, while tariffs on intra-regional trade have been Several initiatives are trying to address this challenge, including formally eliminated by parties to the CMP, measures with Kenyan start-up OkHi. Founded in 2014, OkHi’s aim is to develop equivalent effect remain, including an upsurge in nontariff a physical address system for billions of people globally that are barriers.78 These would need to be removed for an SDM without a physical address, by providing a ‘tag’ for each house to emerge and to increase regional trade with the rest of comprising GPS coordinates and a picture of the house.83 the world. Poor logistics has been a challenge for e-commerce firms The NCIP has also been taking steps to create a Single currently operating in the region and is seen as a barrier to those Customs Territory, using digital technology to reduce who might otherwise be interested in investing. However, the customs delays and improve the free movement of goods. scale offered by an SDM would enable firms to establish regional Objectives include migrating physical controls onto an distribution hubs, reducing the cost and time of deliveries, thus electronic clearance processing system and harmonizing creating a much more effective and vibrant e-commerce market the implementation of a Common Electronic Cargo Tracking and more attractive investment opportunities. This would also System.79 The digitization and standardization of customs reduce reliance on intermediaries, such as EasyBuy, which place and revenue authorities’ information systems has proved orders with international e-commerce companies and then make to be the critical first step, which again suggests that deliveries to customers within their country, often in exchange e-government and the digitization of public service delivery, for cash-on-delivery to overcome payment barriers.84 Regional is an important building block for the development of an scale and seamless logistics, provided through the SDM, would SDM.80 Meanwhile, the AU recently announced the launch also benefit many players in the logistics market, including of the African Continental Free Trade Area (AfCFTA), which distribution companies such as Sokowatch, who are expanding will seek to create a single market for goods and services, their operations in the region.85 with free movement of business persons and investments, across the continent—which could boost intra-African Innovative digital delivery solutions can also help overcome trade by up to 52 percent by 2022—by removing tariff some of the persistent challenges associated with poor transport and nontariff barriers.81 While these developments are and distribution networks. Unmanned drones are being piloted by very encouraging, and significant progress has been made, both the private and public sector to increase access and delivery further efforts in this field are needed to remove cross- of good and services. The Rwandan government is, for instance, border barriers to trade (both digital and analog). trialing drones to support the distribution of medical supplies to remote areas. Global e-commerce giants such as Amazon are also Cross-border deliveries of consumer goods, purchased exploring drones as options for parcel delivery, as are regional through online e-commerce services, also require robust firms such as Mobisol.86 and integrated logistics systems, able to deliver goods to and from the border, and on to customer premises or Leveraging the region’s appetite for innovation—a track record shared local delivery points. These systems are currently for embracing challenges and turning them into opportunities— weak due to the trade and customs barriers detailed earlier, gives the region a comparative edge, which could be amplified by and the absence of companies that ship goods across the creation of an SDM and supported by coordinating regional borders. However, the lack of formal physical address policy on emerging technology. systems outside of major cities is also a major barrier.82 78 See http://documents.worldbank.org/curated/en/799871468194049251/ pdf/938430WP0v20Bo0arket0Scorecard02014.pdf. 83 See http://disrupt-africa.com/2017/12/kenyan-addressing-startup-okhi-hits- 100k-deliveries/. 79 See https://www.nciprojects.org/project/single-customs-territory. 84 See https://www.easybuyafrica.com/. 80 Based on stakeholder consultation in Kenya. 85 See http://disrupt-africa.com/2017/05/kenyan-merchant-solution-sokowatch- 81 See https://www.aljazeera.com/news/2018/03/african-continental-free- starts-east-african-expansion/. trade-area-afcfta-180317191954318.html. 86 See http://www.dw.com/en/in-africa-get-your-next-delivery-by- 82 According to stakeholder feedback. drone/a-19296585. A SINGLE DIGITAL MARKET FOR EAST AFRICA 31 3.4 Key enablers consumers to charge devices, such as smartphones. Across East Africa, 118 million people still lack access to electricity88—a Many enabling factors contribute to the development and mere 7.6 percent have access to energy in Burundi (as shown in success of critical attributes of the market layers detailed earlier. Figure 13). For example, the expansion of connectivity will be contingent on Figure 13: Access to electricity, 2016 sustainable energy provision. Equally, for IT start-ups to flourish, they need both the human capital to sustain them and a business 100 environment that is conducive to growth. Some enablers that are specific to the SDM, such as bridging/remedying/fixing the % OF POPULATION digital skills deficit, will be directly addressed through the Digital Roadmap. However, not all enablers can be addressed as part 56 of the road map proposed, as they are often broader in nature 50 and concern a much wider audience of stakeholders. These areas 32.8 will, nevertheless, benefit from coordination and parallel efforts 29.4 26.7 to address them—here the SDM roadmap provides a further platform for advocating for concerted action in these key areas. 7.6 8.9 0 For example, improving transport infrastructure is not just critical BURUNDI RWANDA SOUTH SUDAN TANZANIA UGANDA KENYA to economic development, but an important aspect of the overall SDM Vision, particularly in the context of enabling logistics for e-commerce development. Though addressing major transport infrastructure deficits is beyond the direct scope of the Digital Source: World Bank, 2016.89 Roadmap, close coordination with the transportation sectors is In other words, many more citizens in East Africa have access to encouraged, not least in view of the potential efficiency gains mobile networks than to the electricity needed to charge their to be made (for example, by coordination network deployment devices. However, digital platforms and solutions can be leveraged with roadworks). Considerable cost savings of up to 70 percent to expand access to power. Mobile-money-enabled off-grid to 90 percent are possible when fiber is installed alongside other alternatives, such as M-KOPA, Mobisol, and BBOXX are distributing infrastructure.87 Many projects to improve highway and railway household solar power solutions to over 1 million homes across infrastructure are under way in the region. The Northern Corridor Kenya, Tanzania, and Uganda.90 Some of these companies have program is, for example, helping to fast-track the development even started adding mobile connectivity and data plans to their of a standard-gauge railway linking Rwanda and Uganda to the solar power bundles. Yet, their market expansion is hampered by port of Mombasa. These projects will not only help facilitate the limited interoperability of existing digital payment systems regional goods trade but also provide significant opportunities and the absence of harmonized data privacy legislation (noted for encouraging cross-border fiber network links at low earlier), illustrating the need to fully implement a single data additional cost. and online market. Though not an ‘online service’, continuing to Similarly, there are synergies between access to energy and leverage existing synergies, between expanded access to ICT and connectivity. The availability of power is a significant barrier to energy will be critical to accelerating progress toward an SDM. the development of the digital economy in East Africa. The poor Investment in telecom infrastructure can, for example, benefit coverage and reliability of existing power networks increases from coordination with utility networks. Kenya Power has, for reliance on alternative power sources, such as diesel generators, instance, helped deploy over 4,000 km of fiber optic cable along which in turn drive up the operational costs associated with its transmission lines.91 running telecom networks and data centers. The low coverage 88 Based on calculations from https://trackingsdg7.esmap.org/. of grid power also makes it more difficult and expensive for 89 See https://data.worldbank.org/indicator/SP.POP.TOTL?locations=ID or https:// trackingsdg7.esmap.org/. 87 See http://www.worldbank.org/en/topic/ict/brief/making-rural-broadband- 90 See http://solar.m-kopa.com/about/. affordable. 91 See http://www.techweez.com/2016/10/28/safaricom-pilots-fiber-home-nairobi- aims-zukus-jugular/. 32 A SINGLE DIGITAL MARKET FOR EAST AFRICA A conducive regional and national business environment is An important part of the business environment is taxation critical for the creation of digital companies—enabling innovative policy,92which is likely to affect entrepreneurship, innovation, enterprise to grow locally, regionally, and ultimately globally— profitability, and choices concerning where to locate operations but also to encourage innovation among established businesses. of digital firms within the region. For93example, reports suggest Deployment of connectivity infrastructure and the development that taxes levied on tech start-ups in Kenya can be prohibitive.94 of data-driven and online services will first and foremost be As noted earlier, high taxes on telecom equipment and services private sector driven. While movement toward an SDM will create are suppressing investment in rural areas and consumer demand a much larger regional market that benefits entrepreneurs—by across the region. Discrepancies in tax policy, including tax expanding the size of the possible customer base and helping rebates to attract companies, can also cause friction between highlight success stories—concerted action at both the national countries. For example, such incidents in the European Single and regional level is needed to implement policies that encourage Market have triggered reviews of related policy frameworks.95 innovation and increase access to capital. Any tax regime adopted in East Africa should thus balance the need to generate revenue and attract business, with the need The existing business regulatory environment is hampering the to avoid distortions to regional taxation policy. East Africa can ease of doing business in East Africa, and thereby affecting the learn from the EU and should strive to adopt international best creation of a potential SDM. In Uganda, for example, there are practice. The Multilateral Competent Authority Agreement, reportedly 13 separate procedures required to start a business.92 introduced by the Organisation for Economic Co-operation and Most East African countries rank poorly in a corresponding World Development (OECD) in 2016, advocates for automatic sharing of Bank index, with Rwanda ranking the highest and South Sudan country-by-country information on the revenues and taxation of the lowest in the region (see Figure 14). These challenges are multinational corporations.96 compounded as digital firms, particularly those in e-commerce or sectors with licensing or certification requirements, seek to expand regionally, and must navigate these complex barriers 92 See http://reports.weforum.org/global-information-technology-report-2016/ networked-readiness-index/. to establish themselves in each country rather than expanding 93 See http://www.doingbusiness.org/rankings?region=sub-saharan-africa. seamlessly across borders. While large multinational firms have 94 Based on consultations feedback from Kenya. the resources to engage in this costly and time-consuming 95 Companies raising concerns in Europe include not only Apple, Google, and process and can tackle the registration process in countries Amazon but also Starbucks and Fiat. https://www.theguardian.com/world/2016/ aug/30/after-apple-the-other-tax-deals-in-the-european-commissions-sights. simultaneously, local start-ups often do not, putting them at 96 See http://www.oecd.org/newsroom/a-boost-to-transparency-in-international- a disadvantage. tax-matters-31-countries-sign-tax-co-operation-agreement.htm. Figure 14: Ease of doing business index, 2016–2017 Country Global ranking (out of 190) 2017 Sub-Saharan Africa ranking (out of 48) 2016 Rwanda 56 41 2 Kenya 92 80 3 Uganda 115 122 12 Tanzania 132 137 15 Burundi 157 164 34 South Sudan 186 187 46 Source: World Bank, June 2017.92 A SINGLE DIGITAL MARKET FOR EAST AFRICA 33 Complexities also arise where, for example, online services are hosted in one country but served in another or when “ infrastructure is deployed between countries—creating potential Technology and innovation for duplication of taxation in multiple countries. A recent example hubs, incubators, and includes the dispute between the Kenyan Revenue Authority (KRA) and the Wananchi Group, which offers satellite content, over its accelerators are key in tax status. The KRA has argued that while the Wananchi Group is supporting this innovation registered in Mauritius, it had managed some if its services from ecosystem and are becoming Kenya. However, the Wananchi Group claimed its operations were independent and should be separately audited.97 increasingly common across East Africa has also seen the rapid emergence of a vibrant innovation and start-up ecosystem. Start-ups are key sources of innovation and due to their agility are better positioned to be East Africa. ” ahead of the curve in technology adoption and adaptation. They In 2016, of some 173 tech hubs and incubators located across could be a key resource to revolutionize traditional brick and the continent, 35 could be found in East Africa.99 However, over a mortar industries across East Africa, which are currently losing two-year period, this figure has doubled across Africa and in the their competitiveness. region—30 hubs are now estimated to be located in Kenya alone, and 16 hubs are located in Uganda, which feature a number of Technology and innovation hubs, incubators, and accelerators large start-ups.100 Uganda tech and innovation space, HiveColab, are key in supporting this innovation ecosystem and are becoming is currently supporting tech start-ups in the financial, education, increasingly common across East Africa. Start-up accelerators and agriculture sectors, including the agricultural information support early-stage, growth-driven companies through education, app mFarm.101 KLab in Kigali is providing an open space for mentorship, and/or financing. Start-ups enter accelerators for IT entrepreneurs and regularly hosts workshops, bootcamps, a fixed period and as part of a cohort of companies. Start-up hackathons, and networking sessions. KLAb success stories incubators support the creation and initial growth of new and include Academic Bridge, a tech education start-up, and AirClerk early-stage enterprises through access to resources such as capital, a cashless payment system. 102 physical space, networking connections, and mentorship. Their support can last for a longer period than the support offered by Currently, many such entities operate on a national level; however, accelerators. The period can be at minimum six months; however, the most successful are starting to draw participants from across some incubators’ support lasts longer. They offer a pathway to the subregion. These networks are providing a supportive the first investment for start-ups. Rapid technical skills training breeding ground for start-ups in the region. The creation of an providers or technology bootcamps offer short-term, applied, SDM in East Africa could accelerate the development of these intensive technology skills training paired with collaborative regional networks and collaborations, which could in turn speed problem-solving and other soft skills development. These entities up technology adoption and adaptation across East Africa and provide a space where entrepreneurs can go online, share ideas, beyond. mFarm was, for example, originally incubated by the develop skills, and meet potential investors and partners. They mLab in Kenya and is now spreading to other countries in the often also help companies navigate the complex regulatory and region—it is thus a pertinent example of the potential of regionally legal environment at both the national and regional level. In the networked tech hubs to support technology diffusion. Regional past, innovation hubs have successfully facilitated the creation of programs could also encourage more female entrepreneurship. new apps and services that respond directly to local needs and consumer demand. 98 99 See http://blogs.worldbank.org/ic4d/importance-mapping-tech-hubs-africa-and- beyond; figures from 2016. 97 See http://www.techweez.com/2016/08/16/kra-zuku-3-4-billion/. 100 See https://www.gsma.com/mobilefordevelopment/programme/ecosystem- 98 See http://www.infodev.org/infodev-files/mobile_outcome_ accelerator/africa-a-look-at-the-442-active-tech-hubs-of-the-continent/. assessment_02-06-2014_last_version_1.pdf. 101 See https://hivecolab.org; https://thenextweb.com/africa/2011/05/31/why- nairobi-is-exploding-as-the-tech-hub-of-east-africa-interview-with-erik-hersman/. 102 See https://klab.rw. 34 A SINGLE DIGITAL MARKET FOR EAST AFRICA Figure 15: Tech hubs and incubators by country market could also help drive improvements in the quality and targeted support offered to clients of such entities. All these 40 benefits have been recognized by AfriLabs, which leverages a network of some 50 technology and innovation hubs, across 2016 more than 20 countries in Africa, to support new enterprises.104 2018 Some countries are consciously supporting business innovation 30 30 through public programs. The Uganda National Council for Science and Technology, for example, seeks to offer innovative leadership in the development and promote the application of science and technology and its integration in sustainable national development.105 Greater regional coordination between 20 innovation agencies could again reap the benefits offered by a 16 16 larger network, while promoting best practice across the region.106 Access to finance and investment are critical factors affecting 10 the growth of innovative start-ups in the region, as well as the ability of established companies to expand into regional and international markets. Many companies in Africa find it 4 7 6 challenging to obtain debt and equity financing to support their 1 1 growth, due to both high collateral requirements and interest 0 rates. The former is impractical for many start-ups and SMEs. The basic lending interest rate in Kenya was 14.8 percent in Burundi, BURUNDI RWANDA SOUTH SUDAN TANZANIA UGANDA KENYA 17 percent in Kenya, and 21.3 percent in Uganda 2016, whereas most developed economies typically offer similar loans at low single-digit rates.107 Source: World Bank 2016; GSMA 2018.103 The level of venture capital in the region is still low compared to the rest of the world, though the situation is improving. The Note: *2018 data: 1–4 in Burundi, Rwanda, and South Sudan; 5–9 in Tanzania. Global Competitiveness Index from the World Economic Forum However, more can be done to support the regional acceleration (WEF) rates venture-capital availability in East Africa. In this index, of start-ups as well as established technology companies with Uganda and Rwanda scored 2.4 and 3.4. out of 7, respectively, high growth potential. The expansion of these companies could, compared to the United States, which tops the ranking with a for example, be boosted through a coordinated regional program score of 5.2 out of 7.108 Historically, venture capital and private that includes acceleration support for scale-up of proven startup equity firms have tended to fund more established businesses concepts which is often lacking in the region. over start-ups. However, new investors are emerging in the region. Firms like Savannah Fund are providing seed capital to By fostering the development of more regional incubation and high-growth technology start-ups in East Africa, which cannot acceleration programs, East Africa could also become a more- yet attract venture-capital investment.109 Several ‘angel investor’ attractive destination for venture capitalists and investments from multi-nationals, offering a larger pipeline of promising start-ups. A 104 See http://disrupt-africa.com/2017/06/afrilabs-welcomes-new-hubs-into-pan- african-network/. more integrated network would also help to internationalize the 105 See https://uncst.go.ug/who-we-are/. talent-, coach- and mentor-base across the region, which could 106 See https://elibrary.acbfpact.org/acbf/collect/acbf/index/assoc/HASH01ad/ strengthen start-up teams and collaboration, leading to enhanced e44e7241/b749d69a/1a6c.dir/ACR2017%20English.pdf. product and business development across the continent. A larger 107 See https://data.worldbank.org/indicator/FR.INR.LEND. 103 Figures from 2016. See http://blogs.worldbank.org/ic4d/importance-mapping- 108 WEF Global Competitiveness Index, 2018, Venture capital availability: Burundi tech-hubs-africa-and-beyond. Figures from 2018. See https://www.gsma.com/ (2.6), Kenya (2.9), Rwanda (3.4), Tanzania (2.4), Uganda (2.4), and South Sudan mobilefordevelopment/programme/ecosystem-accelerator/africa-a-look-at-the- (no data). 442-active-tech-hubs-of-the-continent/. 109 See http://savannah.vc/about/. A SINGLE DIGITAL MARKET FOR EAST AFRICA 35 networks also exist in the region, providing funding to early-stage Figure 16: Adult literacy rate, 2015 start-ups. Examples include Intellecap, which launched an East African chapter of its angel network in 2015.110 Another growing 100 source of investment is crowdfunding, which is coordinated by organizations such as Jumpstart Africa.111 Moreover, capital is 85% being sourced through grants and competitions. In 2017, Ecobank ran a ‘Fintech Challenge’ event, awarding funding and expert 80 80% mentoring to winning proposals.112 78% 74% 71% Local capital markets are also growing, offering investment in local currencies that limits currency-exchange risks. Local investors 60 will also often have a better understanding of the local market, allowing them to make more informed investment decisions. A technology project investment fund (TPIF) is being spearheaded by the Northern Corridor Technology Alliance (NCTA), which will 40 source financing from local banks and equity pools. 32% The scale offered by the SDM should make it easier for many businesses to attract debt or equity financing, by demonstrating greater growth potential, diverse market exposure, and spreading 20 investment risks. This should also attract greater levels of foreign direct investment (FDI). However, stronger SDM branding could also help promote the potential of East Africa to global investors. 0 An additional factor constraining the development of the regional BURUNDI RWANDA SOUTH SUDAN TANZANIA UGANDA KENYA digital economy is the persistent digital skills deficit, both in terms of the basic digital literacy that consumers require to use digital services and the availability of the more advanced ICT skills Source: World Bank 2017. needed to support development of the tech industry. Large parts of the population remain offline. A lack of familiarity and training All countries in the region currently implement a range of domestic is often cited as a key barrier to adoption; lack of digital skills was initiatives to promote basic digital literacy, including programs cited as the main reason for not going online by 45 percent of to deliver ICT education in schools. Yet teachers often lack the those surveyed in Tanzania and 37 percent of those surveyed in training to fully incorporate ICT in education. Schools also lack Kenya, in a recent GSMA survey.113 Gaining basic digital skills is internet connections and students lack access to digital devices. an even greater challenge for those who also lack basic literacy, In 2016, the Rwandan government approved the inclusion of which is estimated to be roughly 15–30 percent of the population ICT training in their national education policy, with the aim of in most East African countries and far higher in South Sudan (see supporting the implementation of ‘smart classrooms’ in all schools Figure 16)114. by 2019.115 Regional collaboration could leverage examples of best practice in the region and make the development of shared 110 See http://disrupt-africa.com/2015/10/intellecap-launches-angel-investment- educational resources more cost-effective. network-in-east-africa/. 111 See https://www.indiegogo.com/projects/empower-africa-s-entrepreneurs-let-s- Community hubs can also become centers for learning, where jumpstart-africa#/. 112 See http://disrupt-africa.com/2017/05/finalists-announced-for-ecobank-fintech- training to those above school age can be provided. In some challenge/. cases, these centers can also provide the only affordable access 113 See https://www.gsmaintelligence.com/ to connectivity available, adopting a free-usage or shared-access research/?file=8170bf058e42cdb8c186c6c75fb2b30e&download. 114 Latest figures available, from 2015. See https://data.worldbank.org/indicator/ 115 See http://mineduc.gov.rw/fileadmin/user_upload/pdf_files/ICT_in_Education_ SE.ADT.LITR.ZS. Policy_approved.pdf. 36 A SINGLE DIGITAL MARKET FOR EAST AFRICA model. One example is the Maarifa Centers, established by the implementing the initiative, currently lacks funding. Equally, more Arid Lands Information Network (ALIN) across Kenya, Uganda, and resources will be required to support harmonization of teaching Tanzania, which train local communities in basic digital skills that standards at the national level.121 are directly relevant to them—such as how to use mobile phone applications to improve incomes from agricultural production.116 The free movement of students and skilled professionals in the IT Another example is UTouch, which operates ‘digital centers’ in field is another important input to the regional digital economy. remote villages in Uganda.117 Alongside more cooperation on training, greater harmonization of labor laws and training across the region could help both As the SDM develops, so should the demand for basic digital skills identify and fill skills gaps in the region. Work-permit restriction training. As new emerging online services and local content is has seen countries such as Tanzania experience loss of skilled developed, the internet should become more attractive to non- employees and entrepreneurs to other countries in the region.122 users and more attuned to their needs. However, more advanced Differences in labor regulations across the regional market can digital skills will be required by those developing said services and also create arbitrage. content. A 2015 survey found that Kenyans prefer downloading foreign apps due to the poor quality of local apps and the lack of Meanwhile, companies in the region reportedly still struggle to original or useful content.118 Building a world-class ICT workforce, find the technology talent they seek, which suggests that both with advanced digital skills, is therefore necessary to boost local the quality and availability of existing training needs to improve. development of high-quality technology products. Moreover, In the interim, this skills gap is being bridged through short-term, many jobs associated with the fourth Industrial Revolution, noted immersive, and applied training programs, or so-called ‘boot earlier, will require highly specialized skills in science, technology, camps’, that often aim to improve participants’ coding skills. Some engineering, and mathematics. initiatives also specifically target women and girls. Prominent examples include Nairobits that offers training in coding and A growing number of colleges and universities offer ICT-related web design to women.123 Closer public-private collaboration courses. Specialist technical institutions are also emerging, such in developing new training programs can also both tailor skills as Carnegie Mellon University in Kigali 119 or the African Advanced generated to industry needs and maximize the resources available Level Telecommunications Institute (AFRALTI), which is active to digital education. across Kenya, Tanzania, and Uganda.120 The World Bank is also harnessing economies of scale by funding the creation of regional Ultimately, the policy coordination and harmonization of laws centers of excellence for technical training, across Southern and and regulation needed to move toward an SDM will require bold Eastern Africa, which includes centers in Kenya, Rwanda, Tanzania, leadership. While many regional institutions are actively working and Uganda. in this space (detailed in Annex C), coordination and capacity building are needed to support the successful implementation These initiatives will also benefit from ongoing efforts to of the initiative, avoid duplication, and maximize available harmonize the higher education and training systems in the resources. The SDM initiative proposed should thus place a strong region. The EAC endorsed the establishment of a Common Higher emphasis on working through regional institutions, capitalizing Education Area (CHEA) in May 2017, which will promote common on momentum created by existing integration efforts and on standards across the region and enable the mutual recognition providing technical assistance. of qualifications to facilitate labor mobility across the region. However, the Inter-University Council of East Africa, charged with 116 See http://www3.weforum.org/docs/WEF_Internet_for_All_Framework_ Accelerating_Internet_Access_Adoption_report_2016.pdf. 117 See http://u-touch.org/. 118 See http://buzzkenya.com/mobile-phone-users-kenya-prefer-foreign-apps- 121 See http://iucea.org/eahea1/declaration/. kenyan-made-apps/. 122 See http://pwc.blogs.com/legal/2016/09/new-requirements-for-work-permit- 119 See http://www.cmu.edu/africa/. applicants-following-under-new-immigration-law-tanzania.html. 120 See http://www.afralti.org/. 123 See http://www.nairobits.com/about/. A SINGLE DIGITAL MARKET FOR EAST AFRICA 37 4. Conclusions and next steps toward implementation implementation. It also outlines the key regional institutions and ongoing regional initiatives that will be leveraged to achieve the “ The Digital Roadmap SDM objectives (Annex C). Finally, the detailed economic analysis estimating the significant boost to GDP growth and job creation is expected to provide as a result of implementing the SDM is included as Annex D. the basis of coordinated By taking proactive action and working together rather than in competition, the result can be mutually beneficial for all East efforts to implement the African countries, whether they are more or less advanced in their current digital development. While countries that score ” recommendations. higher in their digital scorecard can anticipate increased market access and options for services, countries that present smaller markets and or are currently scoring lower will expect to see benefits from increased investment, services, and infrastructure. Achieving an SDM would have tremendous benefits for East Most importantly, the SDM can help close the ‘digital divide’ Africa’s citizens, governments, and companies—increasing the among East Africa’s people, ensuring that everyone has access region’s competitiveness, growth, and job creation and enabling to and is benefitting from digital technologies rather than just a it to excel in the economy of the future. However, getting there privileged few. will not be easy. Realizing this vision will require a long-term strategic perspective, bold leadership, and, at times, a willingness Looking ahead, the Digital Roadmap is expected to provide to sacrifice elements of national sovereignty and short-term the basis of coordinated efforts by East African governments, comparative advantage for the sake of a much larger but possibly regional institutions, donor partners, NGOs, and private sector to more distant reward for all. implement the recommendations. The World Bank Group, as a neutral broker of knowledge and expertise and as a multilateral This report presented a vision for an SDM and an assessment financing institution, is well positioned and prepared to contribute of the strengths, weaknesses, and opportunities that must be toward this effort, leveraging its instruments for investment and addressed and captured to achieve an SDM. The annexes that technical advisory support as well as its convening power to follow present a detailed roadmap for achieving that vision (Annex bring together relevant stakeholders and draw attention to this A) along with a scorecard (Annex B) to measure progress toward important agenda. 38 A SINGLE DIGITAL MARKET FOR EAST AFRICA Annex A: SDM Roadmap The SDM Roadmap details the recommended policy reforms and completion time frame: (a) short term (1–2 years), (b) medium investments corresponding to each market layer of the strategic term (2–3 years), and (c) long term (3–5 years). Each action has framework and the analysis of current market strengths, also been assigned a priority order (1st or 2nd order priorities) weaknesses, and opportunities in Section 3. Each action indicating the relative importance/impact. identified has been assigned an estimated implementation SINGLE CONNECTIVITY MARKET Action Level Time Frame Priority Order Objective: Stimulating connectivity infrastructure development Undertake PPP investments in regional fiber backbones to increase capacity, Regional Medium–long 1st redundancy, and competition between networks, including opportunities for co-deployment with regional linear infrastructure (roads, pipelines, power transmission). Coordinate national PPP investments to promote universal broadband access National (regional Medium–long 1st in each East African country (last mile connectivity). coordination) Coordinate a reduction in taxation, fees, and procedures for infrastructure deployment: • Reduce or eliminate import duties for network equipment; • Reduce fees and provide easier access to rights-of-way to lay National (regional cables and mount equipment along other public infrastructure Short–medium 1st co-ordination) (such as roadways, electricity transmission lines and streetlight poles, pipelines, rail, and so on); and • Lower cost or free licensing and spectrum allocation for infrastructure and services deployment in rural areas. Develop and implement best practice guidance for the following: • Infrastructure deployment (for example, ducting specifications and use of multiple fiber pairs); Regional • Coordination with other infrastructure projects to encourage (with national Short 2nd parallel fiber deployment (employing a ‘dig once policy’); and implementation) • Regional infrastructure sharing and wholesale access (for example, related to duct access and dark fiber provision). Investigate the feasibility of a single regional 4G or 5G license (including Regional Long 2nd coverage obligations for each country). A SINGLE DIGITAL MARKET FOR EAST AFRICA 39 SINGLE CONNECTIVITY MARKET Action Level Time Frame Priority Order Regional Develop policy to support trials of innovative technologies and business (with national Medium 2nd models for extending connectivity infrastructure and services. implementation) Objective: Lowering regional transit costs for landlocked countries Establish a low-cost, open access regional backbone interconnection regime: • Adopt regional open access policy, including rights for any licensed operator in the region to purchase wholesale transit Regional Medium 1st capacity and access submarine cables on non-discriminatory terms through any country in the region; and • Establish regional glidepath for standardizing and lowering interconnection rates and transit pricing. Regional Ensure that ISP licenses and IXP policies enable regional connectivity and (with national Medium 2nd traffic exchange at any IXP in the region. implementation) Objective: Improving affordability and quality of connectivity services Conduct a regional peer review and benchmarking of essential market competition regulations and policies in each country (including licensing, Regional interconnection, number portability, infrastructure sharing, price regulation (with national Short 1st of dominant market players, and so on), and implement or update policies implementation) where deficient. Coordinate a regional reduction or elimination of import and services taxes for consumer devices and connectivity services (by lowering consumer National costs, reducing tax arbitrage and leakage of devices, and portage of data or (regional Short–medium 1st voice bundles across borders from lowest taxed domiciles, resulting from coordination) ONA implementation). Extend ONA coverage to more countries and services: Regional • Extend ONA for voice to Tanzania and Burundi; and (with national Short 1st • Extend ONA to cover data services in all countries. implementation) 40 A SINGLE DIGITAL MARKET FOR EAST AFRICA SINGLE DATA MARKET Action Level Time Frame Priority Order Objective: Removing restrictions to cross-border data flows, storage, and processing Remove any undue restrictions to the free flow, storage, and processing of National (regional data across countries in East Africa and globally (remove data sovereignty Short 1st coordination) requirements for non-essential data). Review legislation relating to intermediary liability for third-party data to Regional ensure that content created in one country and hosted or made available (with national Medium 2nd in another does not face undue restrictions. implementation) Objective: Improving data privacy and cybersecurity Develop and implement regionally harmonized data protection and Regional privacy laws and regulations, including explicit references to cross-border (with national Medium 1st data flows, definitions of personal and sensitive data, and accounting for emerging services (for example, cloud services, data analytics). implementation) Establish a regional cybersecurity task force, building on that proposed by the NCIP. The task force should include all six countries and aim to: • Drive collaboration on the detection and prevention of cybersecurity incidents; • Implement a regional IT platform over which to share Regional Medium 2nd cybersecurity information; and • Coordinate on developing and delivering national cybersecurity awareness raising and training programs for businesses, government agencies, and citizens. Objective: Increasing access to data for development of digital services and analytics Digitize key government registries using regionally harmonized data standards to enable future cross-border data exchange (for example, National (regional to facilitate one-stop border posts or Know Your Customer (KYC) Medium–long 1st coordination) requirements) and the creation of larger regional data sets essential for data analytics and development of regionally relevant data-based services. Develop and implement a regional open data initiative to make regionally Regional standardized data sets available to public and private sectors to stimulate (with national Medium–Long 2nd data-driven services, analytics, and innovation. implementation) Objective: Promote development of regional data centers and cloud services Encourage the development of high security/performance (tier 3) regional data centers through outsourcing of government data storage and cloud services to private sector and/or the development of shared public- National Short–medium 1st private data center facilities; encourage backup/disaster recovery through data centers in other East African countries. A SINGLE DIGITAL MARKET FOR EAST AFRICA 41 SINGLE ONLINE MARKET Action Level Time Frame Priority Order Objective: Facilitating seamless, low cost, cross-border digital payments Establish full interoperability between mobile money networks (domestically Regional and regionally) with a glidepath toward lower fees for cross-platform and (with national Short–medium 1st cross-border transactions; pursue initiatives to reduce currency exchange fees as interim solution (in the absence of a single currency union). implementation) Support South Sudan in the licensing of mobile money services. National Short 1st National Harmonize laws and regulations that affect the availability and ability to use (regional Medium 2nd ‘traditional’ electronic payment platforms such as Visa and PayPal. coordination) Objective: Enabling cross-border identification and verification for digital transactions and services Establish mutual recognition of national digital IDs and a regional platform National for identification verification by governments and digital services providers (regional Short 2nd to enable cross-border data exchange and to meet KYC/user authentication coordination) requirements for cross-border digital transactions. Objective: Harmonizing and modernizing legal and regulatory frameworks for digital transactions Implement harmonized national e-transactions laws (including recognition National of electronic signatures and harmonized consumer protection rules) in (regional Medium 1st accordance with the EAC Electronic Transactions Act, including for Burundi coordination) and South Sudan. Objective: Increase digitization and improve interoperability of public services and information systems (domestic and regional) Develop an over-arching e-government strategy for the region, including defining regional standards for ICT hardware and software procurement, to Regional Short 2nd ensure interoperability wherever possible. Digitize public services over time, based on regional standards, and work to National Medium 1st integrate with regional services where appropriate. Objective: Spur innovations in logistics to facilitate e-commerce deliveries (domestic and regional)124 Adopt and implement regionally harmonized address systems to enable Short parcel delivery, including official adoption of alternative geospatial grid-based (alternatives); systems that can be rolled out instantaneously (for example, what3words, Regional 2nd OkHi, and similar approaches) to compliment traditional addressing systems long which will take longer to implement and may not be well suited to rural areas. (traditional) Adopt an opportunity-focused (as opposed to risk mitigation focused) approach to regulation of emerging logistics technologies (for example, Regional Long 2nd unmanned cargo drones) to foster regional innovation and support with investments in related infrastructure (such as drone ports). 124 In general, domestic and cross-border logistics constraints are outside the scope of the SDM initiative and will be addressed through parallel efforts at the regional and national level such as through improvements in transport infrastructure, one-stop border posts and efforts to remove nontariff barriers. 42 A SINGLE DIGITAL MARKET FOR EAST AFRICA ENABLING ENVIRONMENT Action Level Time Frame Priority Order Objective: Increase digital literacy and advanced skills Invest in comprehensive digital literacy support programs: Ensure internet connectivity, availability of tablets/computers, digital curriculum, and training resources for teachers at primary and secondary schools; • Support the development of community hubs, where citizens can experience National (with digital technology and receive basic digital skills training (web search, mobile regional peer Medium–long 1st money, social media and productivity software and applications such as word learning) processing, spreadsheets, email, and so on); and • Enable targeted digital skills programming and outreach for marginalized groups such as women, the elderly, the disabled, and low-income households. Implement rapid technology skills trainings such as coding boot camps, with emphasis National (with on inclusion of marginalized groups; target industry/workforce relevant technical skills regional peer Short–medium 1st and general entrepreneurship/business skills. learning) Coordinate efforts to improve collaboration between regional tertiary education institutions on identifying digital skills requirements for the regional workforce, Regional Medium–long 1st developing relevant course materials, and developing appropriate course materials, and coordinating courses between institutions. Establish regional centers of excellence for highly specialized skills sets. Regional Medium 2nd Remove barriers (formal and informal) to the free movement of IT students, National professionals, and entrepreneurs across the region; and implement mutual (regional Medium 2nd recognition of IT qualifications in line with the East Africa Common Higher Education coordination) Area Commitments. Objective: Enable and incentivize digital innovation, entrepreneurship, and content creation Improve the regional support network for digital innovators and entrepreneurs: • Facilitate ‘upskilling’ and networking of national and regional tech hubs, incubators, and accelerators (using a hub and spoke model); Regional and • Cultivate a regional pipeline of start-ups (with a demonstrated proof of concept) Medium 1st national ready to attract international venture capital or angel investors; and • Establish a regional tech entrepreneur association to facilitate government- industry dialogue. Establish a regional body for intellectual property protection: • Develop and implement harmonized regional standards for intellectual property protection; Regional • Ensure all countries are party to major international intellectual property (with national Long 2nd agreements; and implementation) • Consider establishment regional registration and enforcement of intellectual property rights. A SINGLE DIGITAL MARKET FOR EAST AFRICA 43 ENABLING ENVIRONMENT Action Level Time Frame Priority Order Objective: Attract FDI and access to capital for the region’s tech industry Support the development of a ‘Digital East Africa’ brand to attract investment: • Develop a regional investment promotion strategy and marketing campaign; and Regional Short 2nd • Showcase a regional pipeline of start-ups to attract international venture capital and angel investors. Establish mechanisms to increase capital access for regional tech companies, including through: • Regional grant and/or seed funding competitions; Regional and Medium 2nd • Regional equity investment funds and low-cost credit facilities; and national • Tax breaks for start-ups with gradual reversion to standard rates after reaching critical mass. Objective: Leverage technology to overcome hard infrastructure constraints (power, transport)125 Encourage access to energy for the unconnected through technology enabled solutions (as well as policy, regulation, and financial incentives): National (with regional Short–medium 1st • Expand smart metering, and the use of mini-grids; and peer learning) • Explore Pay-as-you-go (called ‘PAYGO’) home solar solutions. Improve cross-border logistics and the utilization of transport assets through digitization, and the regional harmonization of data (including IT standards and greater interoperability): Regional • Support the digitization of key registries and interoperability of critical (with national Medium–long 2nd information systems for cross-border movement of goods and people (including one-stop border posts, citizens authentication, customs, immigration, tax/ implementation) revenue, standards bureaus, and so on); and • Make greater use of electronic cargo tracking. Objective: Strengthen Institutions, Leadership and Decision Making Provide financial and operational support to national governments in developing Regional and and enacting national policy and regulations that align with the Digital Roadmap’s Short 1st national recommendations. Work with regional institutions and national governments and regulatory bodies to improve regional ICT and digital economy statistics collection (including data reliability Regional Medium–long 2nd and the frequency of reporting) to enable improved decision and policy making. Increase the capacity and authority of EACO to serve as a regional regulatory advisor Regional Medium–long 2nd and monitor, in collaboration with national regulators and industry. 125 In general, hard infrastructure constraints are outside the scope of the SDM initiative and will be addressed through parallel efforts at the regional and national level. 44 A SINGLE DIGITAL MARKET FOR EAST AFRICA Annex B: SDM Scorecard Accuracy Many indicators used are also difficult to accurately measure. For example, estimates for internet penetration “ The SDM scorecard reported by the ITU and national telecom regulators tend to vary greatly. The scorecard therefore uses consistent sources across (‘scorecard’), outlined countries, where possible, (for example, the ITU rather than data in this section, provides from individual regulators) to allow like-for-like comparison. a comprehensive Appropriateness. Currently, optimal indicators to track all framework of quantifiable elements of the SDM strategic framework are unavailable. Most existing indicators are nationally focused and do not quantify targets and indicators cross-border barriers, which are at the heart of the SDM Vision. to measure progress Moreover, many are of a qualitative nature. toward achievement of As noted above, coordinated efforts are required to address data the SDM Vision and an gaps, data standardization, and measurement issues, as well as effective means of tracking the development of new more appropriate indicators. Supporting disaggregated data, for example, in terms of gender, will be implementation of the important to ensure that the SDM benefits everyone. The EAC is SDM Roadmap over time, working to improve data collection and reporting in the region, in cooperation with third parties such as the ITU; however, further at both the national and work in this field is necessary. Governments in the region should regional level. ” The scorecard makes use of existing, third-party data which is support this effort and improve the accuracy and timeliness of national indicators. available in the public domain, but there are limitations to this data, Country-level indicators relating to the completeness, accuracy, and the appropriateness of existing indicators. The country-level indicators shown overleaf can be used to compare the current status of each country in East Africa, as well as track the impact of the Digital Roadmap on the development of Data limitations domestic markets over time. A separate set of regional indicators are presented in the following paragraphs. Completeness. Many of the selected indicators have gaps in terms of data availability for several countries. Data points tend Targets. The targets selected reflect ambitious goals for to be missing for South Sudan, both due to its nascent statehood connectivity, which is viewed as foundational to the SDM and weak national statistical capacity. Capacity building related to initiative. The targets selected for data, online services, and digital economy statistics features in the Digital Roadmap to help enabling environment are based on above-average performers in improve data collection going forward. global benchmarking/rankings. 45 In the table below, Red denotes low-end of regional range, Orange denotes mid-range, Amber denotes high-end of range. n/d = no data. Framework Indicator Current Figures Target Category Burundi Kenya Rwanda South Tanzania Uganda Sudan Connectivity 172 138 153 165 152 ICT Development Index - ranking (and score (ICT n/d from 1–10) (ITU, 2017)126 (1.48) (2.91 (2.18) (1.81) (2.19) (7) development) Connectivity 2G network coverage (GSMA, 2017)127 n/d 95 100 n/d 89 n/d 97 (infrastructure) - as % of population (ITU, 2016) 53 94 100 30 95 91 3G network (ITU, 2017)128 40 78 92.1 20 85 64 Connectivity coverage - as % of 90 (infrastructure) (GSMA, 2017) 129 40 80 93.9 20 30.5 45 population 4G network Connectivity coverage - as % of (GSMA, 2017)130 10.7 25 64.3 n/d 27.7 45 75 (infrastructure) population Connectivity Presence of a nationwide fiber backbone Yes Yes Yes No Yes Yes Yes in all (infrastructure) Connectivity Mobile phone penetration - as % based on 48 83.1 69.9 ~21.5 74.4 55.1 90 (services) subscriptions per 100 inhabitants (ITU, 2017) Connectivity Secure internet servers - per 1 million people 1 11 6 0 2 2 15 (infrastructure) (World Bank, 2016)131 Connectivity Number of members of largest IXP 1 30 14 0 36 8 40 (infrastructure) (Packet Clearing House, 2017) Connectivity International internet bandwidth per Internet ~6.1 69.0 7.5 0.4 ~1.7 5.5 90 (services) user (kbit/s) (ITU, 2017) Internet usage (ITU, 2017) Connectivity ~5.2 ~26 ~20 ~6.7 ~13 ~21.9 rates - as % of the (Telecom regulators, 70 (services) 9 90 37 n/d 40 49 population 2016) Mobile-broadband penetration Connectivity - % of active subscribers per 100 inhabitants 8.3 26.2 27 1.1 9.2 33.7 55 (services) (ITU, 2017) Mobile-broadband prices 500 MB - as % GNI Connectivity 22.3 4.4 4.4 14.1 3 10 1 per capita (services) 27.9 4.4 39.2 26.1 5.4 17.5 1.5 (ITU, 2017) 1 GB - as % GNI per capita 126 ICT Development Index 2017. See www.itu.int/net4/ITU-D/idi/2017/. 127 See https://www.gsmaintelligence.com/markets/1887/data/?report=5ad0b91030c4b. 128 Measuring the Information Society Report. Volume 2. ICT Country profiles. June 2017. See https://www.itu.int/en/ITU-D/Statistics/Documents/publications/misr2017/ MISR2017_Volume2.pdf. 129 See https://www.gsmaintelligence.com/markets/1887/data/?report=5ad0b91030c4b. 130 Ibid. 131 Latest data available is from 2016. See https://data.worldbank.org/indicator/IT.NET.SECR.P6 46 A SINGLE DIGITAL MARKET FOR EAST AFRICA Framework Indicator Current Figures Target Category Burundi Kenya Rwanda South Tanzania Uganda Sudan Connectivity Fixed-broadband penetration - as % based on 0 0.3 0.2 0 3.4 0.3 (services) subscribers per 100 inhabitants (ITU, 2017) Connectivity Household penetration of broadband - as % 0.9 2.1 0.2 0 1.9 1.9 30 (services) based on subscribers (Telegeography, 2017) Data (data Data protection and privacy law (UNCTAD, protection and Draft Yes Yes Yes Draft Yes Yes 2018)132 privacy) Data Cyber-crime laws (UNCTAD, 2018)133 Draft Yes Yes Yes Draft Yes Yes (cybercrime) Data Intellectual property protection laws - as score (content from 1 to 7 in Global Competitiveness Index 2.6 3.7 5.3 n/d 3.2 3.3 5.5 regulation) (WEF, 2018)134 Data Open data - as score from 1 to 100 in Open n/d 40.42 19.62 n/d 21.73 11.59 50 (open data) Data Barometer (WWW, 2016)135 Online services Lack of ID ownership - % of total population 27 18 20 53 47 49 5 (ID) which is unregistered (World Bank, 2018)136 Online services Use of digitized ID system No Yes Yes Yes Yes Yes Yes (digital ID) Online services Mobile money usage - as % of adults over the (digital age 15 with mobile money accounts (Findex, 1 73 31 n/d 39 51 90 payments) 2017)137 Online services Usage of digital payments - as % of adults over (digital the age 15 that have made or received digital 4 79 39 7 43 55 85 payments) payments over the past year (Findex, 2017) 138 Online services E-transaction laws (UNCTAD, 2018)139 Draft Yes Yes n/d Yes Yes Yes (e-transaction laws) 132 Online Consumer Protection Legislation Worldwide. 2017. See http://unctad.org/en/Pages/DTL/STI_and_ICTs/ICT4D-Legislation/eCom-Consumer-Protection-Laws.aspx 133 Ibid. 134 The Global Competitiveness Report 2017-2018. See http://www3.weforum.org/docs/GCR2017-2018/05FullReport/TheGlobalCompetitivenessReport2017–2018.pdf. 135 See https://opendatabarometer.org/4thedition. 136 See https://datacatalog.worldbank.org/dataset/identification-development-global-dataset. 137 Latest data from 2017. See https://globalfindex.worldbank.org/ - Burundi data from 2014. 138 Ibid. 139 Online Consumer Protection Legislation Worldwide. 2017. See http://unctad.org/en/Pages/DTL/STI_and_ICTs/ICT4D-Legislation/eCom-Consumer-Protection-Laws.aspx. 47 Framework Indicator Current Figures Target Category Burundi Kenya Rwanda South Tanzania Uganda Sudan Online services (consumer Consumer protection laws (UNCTAD, 2018)140 Draft Yes Yes n/d Draft Yes Yes protection) Online services E-Government - ranking (and score) in global 173 119 138 183 130 128 (digital public survey services) (0.23) (0.42) (0.34) (0.18) (0.35) (0.36) 0.65 (United Nations, 2016)141 Online services Lead time to import - median case as number 15 6 2 (trade and 3 days 3 days n/d 4 days of days (World Bank, 2016)142 days days days customs) Online services Logistics performance - based on global index 2.5 3.3 3.0 n/d 3.0 3.0 4.0 (logistics) (World Bank, 2016, score from 1 to 5)143 Enabling Internet access in schools - as score from 1 to 7 environment 2.5 3.9 4.5 n/d 3.1 3.1 5.5 in Global Competitiveness Index (WEF, 2018)144 (digital skills) Enabling Quality of math and science education - as environment score from 1 to 7 in Global Competitiveness 3.7 4.3 4.4 n/d 2.8 3.1 5.0 (digital skills) Index (WEF, 2018)145 Enabling Local availability of specialized training environment services - as score from 1 to 7 in Global 3 5 4.3 n/d 4 4.3 5.5 (digital skills) Competitiveness Index (WEF, 2018)146 Enabling environment Number of procedure to set up a business - in 3 7 5 n/d 9 13 4 (business & Global Competitiveness Index (WEF, 2018)147 innovation) Enabling Venture capital availability - as score from environment 1 to 7 in Global Competitiveness Index (WEF, 2.6 2.9 3.4 n/d 2.4 2.4 4 (business & 2018)148 innovation) 140 Ibid. 141 UN E-Government Survey. Latest ranking compiled in 2016. See https://publicadministration.un.org/egovkb/en-us/Reports/UN-E-Government-Survey-2016. 142 Latest figures available date from 2016: https://data.worldbank.org/indicator/LP.IMP.DURS.MD?view=chart. 143 Latest ranking compiled in 2016. https://lpi.worldbank.org/international/global. 144 The Global Competitiveness Report 2017–2018. See http://www3.weforum.org/docs/GCR2017-2018/05FullReport/TheGlobalCompetitivenessReport2017–2018.pdf. 145 Ibid. 146 Ibid. 147 Ibid. 148 The Global Competitiveness Report 2017–2018. See http://www3.weforum.org/docs/GCR2017-2018/05FullReport/TheGlobalCompetitivenessReport2017–2018.pdf. 48 A SINGLE DIGITAL MARKET FOR EAST AFRICA Framework Indicator Current Figures Target Category Burundi Kenya Rwanda South Tanzania Uganda Sudan Enabling environment Access to credit - as score from 1 to 100 in 10 75 90 10 65 65 90 (business & Doing Business Survey (World Bank, 2017)149 innovation) Enabling environment Total tax and contribution rate - as % of profit 41.5 37.4 33.2 31.4 44.1 33.8 30 (business & in Doing Business Survey (World Bank, 2017)150 innovation) Enabling environment Ease of doing business - ranking in the Doing Above 164 80 41 187 137 122 (business & Business Survey (World Bank, 2017)151 100 innovation) Enabling Innovation capacity - ranking (and score) 122 80 99 96 102 environment in Global Innovation Index (African Capacity n/d (business & (21.31) (30.95 (27.36) (27.97) (26.97) Building Foundation, 2017)152 innovation) Enabling Quality of overall transportation environment infrastructure - as score from 1 to 7 in Global 2.4 4.3 4.7 n/d 3.6 3.3 5 (hard Competitiveness Index (WEF, 2018)153 infrastructure) Enabling environment Access to electricity - as % of the population 7.6 56 29.4 8.9 32.8 26.7 75 (hard with access (World Bank, 2016)154 infrastructure) Enabling Quality of electricity supply - as score from environment 1 to 7 in Global Competitiveness Index (WEF, 2.4 4.1 4.4 n/d 3.1 3.4 5.5 (hard 2018)155 infrastructure) Enabling environment Capacity - based on Africa Capacity Index (digital 53.4 55.2 68.2 n/d 68.8 54.0 80 (African Capacity Building Foundation, 2016)156 leadership and institutions) 149 See http://www.doingbusiness.org/~/media/WBG/DoingBusiness/Documents/Annual-Reports/English/DB2018-Full-Report.pdf. 150 Ibid. 151 Ibid. 152 See https://www.globalinnovationindex.org/gii-2017-report#. 153 The Global Competitiveness Report 2017–2018. See http://www3.weforum.org/docs/GCR2017-2018/05FullReport/TheGlobalCompetitivenessReport2017–2018.pdf. 154 Latest figures available are from 2016. See https://data.worldbank.org/indicator/EG.ELC.ACCS.ZS. 155 The Global Competitiveness Report 2017–2018. See http://www3.weforum.org/docs/GCR2017-2018/05FullReport/TheGlobalCompetitivenessReport2017–2018.pdf. 156 Latest figures available from 2016: https://www.acbf-pact.org/our-work/how-we-do-it/knowledge-learning/africa-capacity-indicators. 49 Regional indicators The regional indicators listed below reflect cross-border barriers time, as new and better data becomes available. Some indicators at each layer of the SDM that need to be eliminated. Targets have have thus been marked as ‘to be confirmed’ in this iteration of the been set against indicators, illustrating the progress required to scorecard. Regional indicators are not defined for the enabling achieve the SDM Vision. This scorecard is designed to evolve over environment. In the table below, Red denotes low-end of regional range, Orange denotes mid-range, Amber denotes high-end of range. n/d = no data. Framework Category Indicator Burundi Kenya Rwanda South Tanzania Uganda Target Sudan Cross-border Connectivity interconnection of No No No No No No Yes in all (infrastructure) network infrastructure International access to Full Full Full n/a Partial157 Full Full for all Connectivity (services) national backbone countries 3G network (ITU, 40 78 92.1 20 85 64 Above 90% in all Connectivity coverage - as % 2017)158 40 80 93.9 20 30.5 45 countries (infrastructure) of population (GSMA, 2017)159 Harmonized sector policy No No No No No No Yes, based on Connectivity on taxation, spectrum international best (infrastructure) allocation, licensing, and practice rights-of-way Price premium of 500 n/d 293 0 1355 442 344 No more than 50 MB data bundle over Connectivity (services) cheapest country for prepaid data - as % differential Data and SMS roaming No No No No No No Completing the Connectivity (services) costs eliminated ‘ONA’ for East Africa National data No No No No No No No localization Data localization law requirements Data protection and Data (data protection privacy law adoption None Draft None n/d Draft Draft Yes in all and privacy) (UNCTAD, 2018)160 157 Capacity can only be bought through ISPs licenced in Tanzania. 158 Measuring the Information Society Report. Volume 2. ICT Country profiles. June 2017. See https://www.itu.int/en/ITU-D/Statistics/Documents/publications/misr2017/ MISR2017_Volume2.pdf. 159 See https://www.gsmaintelligence.com/markets/1887/data/?report=5ad0b91030c4b. 160 Online Consumer Protection Legislation Worldwide. 2017. See http://unctad.org/en/Pages/DTL/STI_and_ICTs/ICT4D-Legislation/eCom-Consumer-Protection-Laws.aspx. 50 A SINGLE DIGITAL MARKET FOR EAST AFRICA Framework Category Indicator Burundi Kenya Rwanda South Tanzania Uganda Target Sudan Cyber-crime law Draft Yes Yes Yes Draft Yes Data (cybercrime) adoption Yes in all (UNCTAD, 2018)161 Regional harmonization No No No No No No Yes, based on Data (cyberlaws) of data policies and international best protocols practice/ AU law Adoption of Online services (Digital In In Digital ID schemes interoperable Digital IDs Yes Yes In progress Yes ID) progress progress in all countries (national) Travel Travel Travel Online verification only Online services (Digital Regional recognition of only for online services No (not only (not No No ID) digital IDs (not across the region digital digital ID) ID) digital ID) of digital ID Cross-border mobile No No No No No No Yes in all Online services (Digital money platform payments) integration and interoperability Regional cross-border Online services (Digital Premiums transaction premiums No No No No No No payments) eliminated eliminated e-transaction law Online services adoption Draft Yes Yes n/d Yes Yes Yes in all (e-transaction laws) (UNCTAD, 2018)162 Online services Regional harmonization No No No No No No Yes, based on (e-transaction laws) of e-transaction laws international best practice Consumer protection Online services law adoption (UNCTAD, Draft Law Law n/d Draft Law Yes in all (consumer protection) 2018)163 161 Ibid. 162 Ibid. 163 Ibid. 51 In the table below, Red denotes low-end of regional range, Orange denotes mid-range, Amber denotes high-end of range. n/d = no data. Framework Category Indicator Burundi Kenya Rwanda South Tanzania Uganda Target Sudan Regional harmonization Yes, based on Online services of consumer protection No No No No No No international (consumer protection) laws best practice Number of regional No more than Online services (trade nontariff barriers - in 5 23 10 n/d 24 16 5 nontariff and customs) EAS common market barriers scorecard164 Use of charges with equivalent effects on Online services (trade regional trade - number 37 34 35 n/d 28 9 None and customs) of charges noted in ECMA common market scorecard165 Trading across borders Online services (trade - as score from 1 to 100 47.02 67.63 72.44 26.19 20.21 62.8 80 and customs) in Doing Business Survey (World Bank, 2017) Number of regional Enabling environment centers of excellence for Data to be collected (digital skills) advanced digital skills development Number of regional entrepreneurship and Enabling environment innovation networks (business and Data to be collected (incubation, acceleration, innovation) financing, and investment promotion) 164 See www.eatradehub.org/east_african_common_market_scorecard_2016_tracking_eac_compliance_in_the_movement_of_capital_services_and_goods - page 14. 165 Ibid. - page 13. 52 A SINGLE DIGITAL MARKET FOR EAST AFRICA Annex C: Existing integration efforts and their role in the East Africa SDM A number of organizations are already working toward the A 2016 paper by the International Growth Centre estimated that the goal of greater integration of East African economies. The EAC’s CMP had led to a 213 percent increase in bilateral trade and Digital Roadmap aims to build on the progress made by these that full implementation of a single market could lead to a similar organizations and to coordinate new activities with existing magnitude of trade gains in the future.167 initiatives. In this annex, we summarize the current roles played by the following organizations: The EAC has programs covering almost all economic sectors, and four of these are directly relevant to the SDM Vision: • EAC • EACO • Infrastructure - including communication links and hard • NCIP infrastructure such as transport, which is part of the enabling • AU environment. • Other international initiatives • Customs - aiming to establish a single customs territory to support regional trade, which can help promote a single East African Community online market for goods. • Education, science, and technology - harmonizing education The EAC is a regional intergovernmental organization whose mission systems and developing centers of excellence in the region, is “to widen and deepen economic, political, social, and cultural which can help encourage skills training to promote increased integration” between the partner states. After South Sudan joined the infrastructure and online services in the region. community in April 2016, there are now six EAC partner states, which • Immigration and labor - supporting the free movement form the focus of the present SDM project. The first major integration of people and labor in the region, which can help promote milestone achieved by the EAC was establishment of the Customs innovation and technical support where needed. Union in 2005, whereby all partner states “agreed to establish The EAC has put policies and guidelines in place for East Africa across free trade (or zero duty imposed) on goods and services amongst a number of these areas. However, feedback from the consultation themselves and agreed on a common external tariff (CET).”166 suggests that the interpretation and application of these policies This was followed by the establishment of the CMP in 2010, which can vary by country. This is illustrated by the EAC’s ‘Common Market is based on seven freedoms and rights: Scorecard Report’, a monitoring tool for the implementation of the CMP, which shows that each country is implementing reforms at a • Free movement of goods different pace.168 • Free movement of persons • Free movement of labor/workers • Right of establishment 167 See http://www.theigc.org/wp-content/uploads/2016/06/Mayer-and-Thoenig- • Right of residence 2016-Working-paper.pdf. • Free movement of services 168 See https://d3n8a8pro7vhmx.cloudfront.net/eatradehub/pages/2893/ • Free movement of capital attachments/original/1481012380/East-Africa-Common-Market-Scorecard-2016. pdf?1481012380. 166 See http://www.eac.int/integration-pillars/customs-union. 53 A major milestone that will help support the SDM was the The activities of EACO are organized under 11 working groups:172 establishment of the EAC Legal Framework for Cyberlaws (Phase 1 of • Policy and Regulatory Harmonization which was published in 2008 and Phase 2 in 2011). This framework is designed to be used for the reform of relevant national laws, • Infrastructure Development, Connectivity and Digital Inclusion reflecting international best practice to facilitate e-commerce. It is • ICT Services and Applications a broad framework, with Phase 1 covering areas such as electronic • Postal Services Development transactions and digital signatures, cybercrime, data protection and • IP Networks, Standards and Cyber Security privacy, and consumer protection. Phase 2 of the framework focuses • Broadcasting Services Development on issues related to intellectual property, competition, taxation, and information security.169 • Telecom Numbering and Internet Resources Planning, Allocations and Governance The EAC will be an essential stakeholder in the implementation of the • Communications Service Pricing and Industry Analysis Digital Roadmap. The organization already plays an important role in • Quality of Service and Consumer Affairs coordinating regional initiatives and in setting standards to promote • Environment and e-Waste Management harmonization of national laws and regulations. However, the EAC has not achieved all the goals set out at its inception. Implementing • Spectrum Management the Digital Roadmap will require a much greater scale of regional EACO’s ongoing initiatives are closely aligned with the goals of the intervention activities in the digital sector. The SDM Digital Roadmap SDM Vision, particularly those related to a single connectivity market. recommends that the capacity of the EAC is developed accordingly, EACO could be a key stakeholder in the implementation of the Digital along with the appropriate resources to do so, so that it can effectively Roadmap for East Africa, especially in areas relating to licensing and drive progress toward the SDM Vision. Of the six countries in the EAC, the use of spectrum, along with other areas where policies would four (Kenya, Rwanda, South Sudan and Uganda) are also involved in need to be implemented across the six countries. EACO would need to the Northern Corridor Integration Projects, discussed below. work closely with the EAC as it sets policies and with the regulators in each of the six countries. As with the EAC, the greater workload would East African require increased capacity building and the appropriate resources. Communications Northern Corridor Organization Integration Projects Established in 2000, the EACO describes itself as “a regional The NCIPs are designed to support the vision of “A Northern organization that brings together national ICT regulators, operators, Corridor that is fully integrated to improve the competitiveness of services providers (in the telecommunication, broadcasting and postal the region in the global market.”173 The NCIPs are supported by the sub-sectors), ICT training institutions and other stakeholders in the heads of state of Kenya, Rwanda, South Sudan, and Uganda, who communication sector” across the EAC countries.170 EACO’s primary attend summit meetings to discuss progress and agree actions. role is to support the integration of the region’s ICT markets, by driving These four principal partner states have recently been joined by the harmonization of ICT policy and regulatory frameworks. Notably, the Democratic Republic of Congo, and the last summit in Nairobi it has developed a draft regional regulatory framework, alongside was also attended by Burundi and Ethiopia.174 Tanzania is not an implementation and monitoring framework to support adoption currently involved in the NCIPs, but it is engaged in a number of across member states.171 EACO is also tasked with recommending parallel and complementary initiatives through the EAC. best practice for a harmonized and converged licensing framework, including guidelines for the management and allocation of spectrum. There are 14 separate NCIPs, each coordinated by one of the participating countries. One of the NCIPs relates specifically to ICT, but most of the projects will have some impact on the SDM 169 See http://unctad.org/meetings/en/Presentation/CII_EM5_P_RAchieng_en.pdf. Vision, as summarized in Table C.1. 170 See http://www.eaco.int/. 172 See http://www.eaco.int/index.php/working-groups-committees. 171 See http://www.eaco.int/docs/WGsReports/EACO_WG01_Record_of_6th_ Meeting_Arusha-November_2016.pdf. 173 See https://www.nciprojects.org/about/about-us. 174 See http://www.newtimes.co.rw/section/read/193595/. 54 A SINGLE DIGITAL MARKET FOR EAST AFRICA Table C.1: The 14 NCIPs and their relevance to the SDM Vision NCIP Coordi- Relevance to SDM Vision nating Country Air Space Management Rwanda • Promoting lower-cost air travel and more routes that can help facilitate a single labor market • Facilitating drone delivery services for e‑commerce in the long term Mutual Peace, Rwanda • Supporting the security and stability required for a successful SDM Security and Defense • Covering cybersecurity as a central element of national security Cooperation • Establishing a Joint Intelligence Center in Nairobi that requires data sharing between governments Single Customs Rwanda • Enabling the seamless flow of goods required for a single e‑commerce market Territory Immigration, Tourism, Rwanda • Supporting the free movement of people for a single labor market (for example, partner states have Trade, Labor and waived work permit fees for citizens of the Northern Corridor) Services (ITTLS) • Facilitating the use of national identity cards as travel documents that can also help support a single online market Land Kenya • Harmonizing the land acquisition process that can help support a single infrastructure market (including rights-of-way and access to sites needed for data centers and other infrastructure) Human Resource Kenya • Building capacity in numerous sectors, including ICT Capacity Building • Supporting a skills audit for NCIPs, the creation of centers of excellence for skills training, and the removal of non-tuition fee barriers for studying in partner states Financing Uganda • Financing of the other NCIPs, including ICT Fast Tracking Political Uganda • Providing an enabling governance structure for implementation of the SDM Vision Federation Commodities Exchange Kenya • Expanding agricultural commodity exchanges that could be a key application for a single online market Power Generation, Kenya • Increasing the generation of power and facilitating power trade between states, required to power Transmission and ICT infrastructure and devices Interconnectivity Crude Oil Pipeline Kenya • Creating potential for coordination with telecom infrastructure rights-of-way Development Oil Refinery Uganda • Facilitating expansion of supporting hard infrastructure Development Standard Gauge Uganda • Supporting the free movement of people and efficient transportation of goods to accelerate trade Railway and services • Creating potential for coordination with telecom infrastructure deployment ICT Infrastructure Uganda • Coordinating ICT infrastructure, which is a critical element of an SDM (discussed further in this section) Refined Petroleum Kenya • Creating potential for coordination with telecom infrastructure deployment Products Pipeline Development 55 While all the NCIPs have an indirect impact on the SDM Vision, the To implement the SDM Vision across all six countries, the Digital ICT infrastructure NCIP is directly applicable to the SDM Vision. It Roadmap recommends considering whether to expand the identifies nine priority areas that should be addressed. These are Northern Corridor ICT program priority initiatives to Tanzania relevant to many areas of the SDM framework, beyond simply ICT and Burundi, potentially in coordination with the EAC. It is infrastructure, as shown in Table C.2. noted, however, that the heads of state of the Northern Corridor countries have not met for over a year, and thus an early step in Table C.2: Priority areas for the ICT the implementation process would be to determine the future of infrastructure NCIP the NCIPs, and how the NCIP and EAC programs can be rationalized to avoid duplication. SDM Vision Priority Areas for the ICT Frame-work Infrastructure NCIP • ICT policy, infrastructure implementation, Broader initiatives in Africa Single and broadband connectivity Beyond East Africa, there are several organizations whose goal is connectivity • Roaming charges and termination rates to increase integration and cooperation among countries on the market • Harmonization of sim-card registration African continent. regimes The African Union is a political union of all 55 nations in Africa, Single data • Cybersecurity which aims to promote integration of the continent while tackling market social, economic, and political problems. The AU’s vision is to Single online • E-services work toward “an integrated, prosperous and peaceful Africa, market driven by its own citizens and representing a dynamic force • Digital migration in global arena.”176 The AU has established various standards Enabling • Mainstreaming of ICT in the integration and workstreams relevant to the SDM. For example, the AU environment projects members adopted a Convention on Cyber Security and Personal • ICT skills and human capital development Data Protection in 2014. To facilitate implementation of the • Support for development of policy and Convention, ‘Privacy and Personal Data Protection Guidelines’ regulatory framework to the Republic of were recently published, detailing the roles and responsibilities South Sudan of key stakeholders involved in data protection.177 Notably, the AU is spearheading the African Continental Free Trade Area Progress has been made against all of the NCIPs, and the (AfCFTA), which was signed into effect by 44 countries, in March outcomes are published online following each NCIP summit.175 2018, at an AU summit in Kigali. Once it is ratified, the AfCFTA The most significant achievement of the ICT project relates to will create a single continental market for goods and services, as the elimination of roaming charges under the East Africa ONA well as a customs union facilitating free movement of capital and roaming program. business. The initiative aims to boost intraregional trade which currently accounts for no more than 15 percent of total trade in While the NCIPs are led by the public sector, the The Internet Africa.178 According to the United Nations Economic Commission & Television Association (NCTA) aims to coordinate the regional on Africa“(UNECA), intra-African trade is likely to increase by response of the ICT private sector. The NCTA consists of business 52.3 percent under the AfCFTA, due to the elimination of import members that contribute to the NCIP project delivery within duties. This figure is set to double upon further removal nontariff the Northern Corridor. Governments in the Northern Corridor barriers.179 A recent report, advocating for the formation of an have agreed in principle on a regional PPP framework to facilitate coordination of effort with the private sector. This is an 176 See https://au.int/en/au-nutshell. encouraging development, which could be expanded to cover the 177 African Union and Internet Society. May 2018. Personal Data Protection Guidelines for Africa A joint initiative of the Internet Society and the Commission broader East Africa SDM. of the African Union. 178 See https://au.int/sites/default/files/pages/32151-file-plenary_2__brief_on_ 175 See http://nciprojects.org/publications. intra_african_trade_and_investments.pdf. 179 See https://www.uneca.org/stories/signing-afcfta-giant-stride-forward- development-africa-eca%E2%80%99s-vera-songwe. 56 A SINGLE DIGITAL MARKET FOR EAST AFRICA African SDM, under the AfCFTA, suggests that many of the planned Tripartite Free Trade Area between the EAC, Southern economic benefits emerging from the creation of an SDM will be Africa Development Community (SADC), and the COMESA. This realized by leveraging technology-based solutions grounded in aims to create the largest single free trade area in Africa, with Cloud Computing Services (CCS).180 a market of over 600 million people, although there have been delays in the agreement being ratified by all countries.183 The African Council of Regulators (ARC) was established in 2017 and now meets on a quarterly basis. The ARC seeks to implement These broader African initiatives share many of the objectives of the ‘One Africa Network’—a roaming agreement that builds on the SDM and can support its development. The East Africa SDM the successful East Africa ONA.181 Moreover, the ARC is working should seek to align with the standards and regulations adopted on issues such as access to ICT for schools and has established by these pan-African bodies. In fact, East Africa is in a position to working groups in the following areas, which are relevant to the build on existing regional ties to implement changes at a faster SDM: pace than other regional blocs in Africa. However, the SDM proposes to target much deeper digital integration than is likely • Telecommunications to be feasible for the other regions in Africa or Africa as a whole. • Connectivity infrastructure development As a result, East Africa may also help lead the way for some of • Numbering and internet resources planning and allocation these broader African initiatives. • Cybersecurity and online privacy • ICT services and applications Other international • Emerging technologies initiatives • Digital inclusion In addition to the national governments and the regional bodies The SMART Africa Alliance is another continent-wide discussed earlier, a range of international organizations are organization that is addressing ICT issues. It is tasked with working on initiatives that support the SDM Vision. As the Digital implementing the SMART Africa Manifesto that was endorsed in Roadmap is implemented, it will be important to work closely 2014 by all African heads of state and the AU. The Alliance is a with these organizations to coordinate initiatives and avoid multi-stakeholder partnership, which includes member states, unnecessary duplication of effort. For example, establishment of multilateral organizations, NGOs, and the private sector. It focuses an SDM is a pillar of the WEF’s Internet for All initiative in East on goals such as increasing access to digital technology and on Africa. Consultation meetings took place in Rwanda and Uganda key enablers, including innovation and capacity building, that are under the auspices of said WEF initiative, and it is hoped that consistent with the SDM Vision. Today, the Alliance includes a wide WEF will continue to play a key role in the implementation of the range of countries across Africa, including four of the six countries SDM Digital Roadmap. in East Africa (not including Burundi and Tanzania). SMART Africa can be an important partner for improving various aspects of the Other sector-specific bodies are also likely to play a role in SDM. Although its remit is broader than East Africa, it can play a supporting the implementation of various aspects of the SDM. For key role in promoting and coordinating the SDM agenda at the example, the Universal Postal Union (UPU) has existing programs continental level. For example, it recently hosted the Transform in areas such as logistics, payments, digital ID, and consumer Africa Summit, in May 2018, which addressed the conference protection. Cooperation with the ITU on regulation, standards, theme ‘Accelerating Africa’s Single Digital Market’.182 and data collection is also merited. The Digital Roadmap also takes existing World Bank programs into account, many of There are currently also efforts to spur greater integration which have objectives that overlap with the SDM Vision. Several between different regional blocs in Africa. One example is the such programs exist, including ID4D and the East Africa Regional Transport, Trade, and Development Facilitation Project.184 180 See Single Digital Market for Africa Report. Transform Africa Summit May 2018. 181 See https://smartafrica.org/events/past-meetings-and-events/article/3rd-council- 183 See http://allafrica.com/stories/201707190091.html. of-african-regulators-conakry-guinee-14-15-decembre-2017. 184 See http://www.worldbank.org/en/news/press-release/2015/06/11/world-bank- 182 See https://smartafrica.org/events/past-meetings-and-events/Transform-Africa- group-approves-500-million-for-eastern-africa-development-corridor. Summit-2018-09-10-May-2018-Kigali-Rwanda. 57 Annex D: Economic • Furthermore, the analysis shows that existing broadband users will also benefit from regional integration, as they will Impact pay less for broadband services and will be able to access a wider variety and higher value of digital content and services, as well as to connect with a wider network of users across the region. Such savings will result in significant consumer benefit Assessment but will not be captured in GDP growth statistics. • Finally, as digital services such as mobile money become interoperable and competitive across the region, the benefits will extend to the bottom of the pyramid as more and more citizens gain access to such services for spending, earning, and producing their own goods and services. These estimates, even in the High Scenario, are conservative, as a more integrated and competitive regional market will further Background and increase innovation, technology adoption, and investment across all acknowledgements sectors over the medium to long term, fueling a reinforcing cycle of productivity gains, growth, and job creation for many years to come. This analysis has been carried out to estimate the economic impact of implementing an SDM in East Africa. This includes Drivers of economic impact assessing the impact of digital market integration on GDP growth and job creation at the macro level, as well as the distribution of under an SDM the expected benefits across income levels, particularly for those at the bottom of the pyramid. At its heart, the SDM initiative is about creating the economies of scale and network effects necessary for the East Africa It has been carried out with joint effort between teams at the World region to be a competitive player in the global digital economy. Bank, Analysys Mason Limited, and individual economists (Neil The economies of scale and network effects generated through Gandal, Edgardo Sepulveda, and Ivan Gonzalez Berenguer Pena). the SDM will have a ripple effect across the digital economy, boosting broadband penetration and ultimately translating While this assessment focuses on the East Africa region specifically, into GDP growth and job creation. the methodology developed can be applied to assess the impact of moving toward an SDM in any region/country grouping. For the purposes of this paper, the latter (that is, job creation) is estimated using the long-studied links between increased Summary of findings broadband penetration and job growth. For GDP growth, an extensive exercise that looks at direct and indirect benefits The analysis that follows shows that the benefits of integration from changes in broadband penetration, as a result of the toward an SDM in East Africa are significant, particularly in the SDM, has been conducted. countries lagging in broadband availability and adoption. In To assess GDP growth, the general framework shown in Figure summary, the paper shows the following: D.1, which considers four different impact pathways, was used. • In the Base Scenario, the average increase in GDP across The four impact pathways are the following: the region is 0.57 percent five years post-integration, or US$0.93 billion, while the average increase in employment 1. The first effect results from the construction of broadband is 2.2 percent, or approximately 1.6 million new jobs. In the networks. In a way similar to any infrastructure project, the High Scenario, GDP increase is in the range of 1.6 percent, or deployment of broadband networks creates jobs and has a US$2.6 billion and employment increase rises to 6.2 percent, multiplier effect throughout the wider economy. or approximately 4.5 million new jobs. In both scenarios, these benefits are based on the increase in mobile broadband 2. The second effect results from the ‘spillover’ externalities, adoption resulting from integration. which have an impact on both enterprises and consumers: 58 A SINGLE DIGITAL MARKET FOR EAST AFRICA Figure D.1: GDP impact general framework Methodology and literature review BROADBAND DEPLOYMENT CONSUMER An extensive literature review was conducted to determine SURPLUS the best methodology to measure GDP growth, jobs creation, and consumer surplus. See Figure D.2 for papers consulted to RESIDENTIAL determine the economic impact of broadband. PENETRATION Writ large, the research and evidence-base reviewed falls into DIRECT HOUSEHOLD five categories: BENEFITS INCOME • Contribution to economic growth (“positive externalities”) ENTERPRISE • Contribution to productivity gains PENETRATION • Contribution to employment and output of broadband deployment (“countercyclical effect”) • Creation of consumer surplus • Improvement of firm efficiencies TOTAL INVESTMENT IN INFRASTRUCTURE FACTOR These five areas use a number of techniques and methodologies, DEPLOYMENT PRODUCTIVITY the most common of which are the following: • Input-Output Analysis • Econometric Analysis CONTRIBUTION • Consumer Surplus TO GDP GROWTH Source: ITU. Figure D.2: Literature review Title Author (Year) • The adoption of broadband within firms leads to a multifactor The Impact of Broadband on Growth and productivity gain, which in turn contributes to growth of GDP. MICUS (2008) Productivity • Residential adoption drives an increase in household The economic impact of broadband on real income. Koutroumpis (2009) growth: A simultaneous approach 3. Beyond these direct benefits, which contribute to GDP Broadband infrastructure and Economic growth, residential users receive a benefit in terms of Czernich (2011) Growth consumer surplus. The Economic Impact of a European Copenhagen Economics This last parameter, while not being captured in the GDP statistics, Digital Simple Market (2010) can be significant, as it may represent benefits in terms of enhanced The Economic Impact of broadband in access to information, entertainment, and public services. Panama ITU (2012) Therefore, following the general GDP framework from the ITU, this paper will examine the economic impact of an SDM in East Measuring (in a time of crisis) the impact Angelo Castaldo, of broadband connections on economic Alessandro Fiorini and Africa, assessing spillover externalities (changes in GDP and jobs) growth: an OECD panel analysis Bernardo Maggi (2016) and consumer surplus. These methodologies have been used to answer two key questions: The paper also considers the added value of an SDM, in terms of the quality and variety of digital services available at a given 1. Does the economic impact of broadband increase with broadband penetration level in an integrated regional market, penetration and can we pinpoint a saturation threshold when compared with an isolated country market. decreasing returns to penetration exist? 59 2. What explains the lagged effect of broadband in the economy? independent variables such as broadband penetration, level of tertiary education, fixed capital investment, and the Some authors have pointed out a potential ‘saturation’ effect, dependent variables (such as GDP or employment growth). meaning that beyond a certain adoption level (not specified, as of This methodology is the most commonly used because it is yet), the effect of broadband on the economy tends to diminish. rare that more than two years of data across each variable These are key elements to consider when assessing broadband’s is available. Given the need to determine the direction of impact in developing countries, which typically have lower causality, it is common to lag the variables by collecting data penetration and supply. for independent variables in year 1 and regressing them against dependent variables in year 2 or after. • Panel data and simultaneous equations are two techniques Input-output analysis that further help econometric analyses determine causation rather than correlation. They are among the most successful Definition: This approach focuses on determining how much techniques that have been employed in the papers that value added and employment is generated through the rollout analyze broadband economic effects. Panel data is a time of high-speed broadband services. The idea behind this model series for multiple geographic areas, (that is, it is both a time is that complex relationships develop between industries, series and a cross-sectional data set). This allows researchers because each sector sources goods and services from other to account for time fixed effects and geographical fixed effects. sectors. Consequently, investments in one sector indirectly trigger • Simultaneous equations are used to deal with endogeneity or demand in other sectors as well. These networked relationships a cycle where factors cause the indicators to change and vice mean that the effect of investment of a broadband rollout versa. This problem is particularly pronounced in the study program is greater than the direct effects would suggest. of broadband’s effect on GDP, GDP per capita, and income. When there is sufficient data, this approach is optimal. Methodology: The estimation of countercyclical effects comprises Limitations: The key disadvantage of the econometric analysis, two steps—the estimation of investment required to fulfill the particularly for the simultaneous equations approach, is the lack targets of the broadband plan and the calculation of resulting of data available, especially in developing countries. economic effects through input-output analysis. Limitations: Input/output tables are not easily available in many countries, including the ones being assessed for this report. Consumer surplus of broadband Definition: The theoretical framework for the measurement of broadband consumer surplus is based on the notion that a Econometric analysis new good (in this case, broadband) provides benefits that are Definition: The available literature reflects a number of additional to the old (dial-up access). It is not as common in the econometric studies to measure broadband’s impact on GDP literature as econometric studies. growth and employment by specifying regression models Methodology: The objective of this methodology is to calculate where GDP growth, employment, and other output metrics are a metric for consumer surplus and net gain in producer revenue a function of broadband deployment and penetration. However, expressed in a single currency for comparability. due to the limited data availability, most studies tend to focus on developed countries. Limitations: The model measuring consumer surplus originated from broadband services presumes a stable demand as core Methodology: There are three types of model-estimation factors shaping demand do not change substantially. In that sense, procedures used to assess the economic impact of broadband. consumer surplus results are quite valid for the short run. On the These include (a) cross-sectional regression, (b) panel data, and other hand, the analysis can yield conservative estimates because (c) simultaneous equations. they might exclude gains to early adopters, shifts in demand linked • The cross-sectional regression relies on one observation per to GDP growth, falling prices of personal computers, greater unit (country, county, region, and so on). When studying change capability of online system, and changing user willingness to pay. in variables, at least two points in time are needed. It includes Furthermore, the methodology excludes indirect benefits. 60 A SINGLE DIGITAL MARKET FOR EAST AFRICA Therefore, the preferred approach to measure SDM impact in East Africa is one that uses panel data econometric analysis, using Macroeconomic impact broadband penetration rates as proxy to evaluate spillover effects In this section, we look at the impact of an SDM on GDP and jobs, (GDP growth and jobs creation), and a separate assessment of the based on an increase in mobile broadband adoption, as well resulting implications in terms of consumer surplus. Broadband as the benefits that will accrue to existing mobile broadband saturation effect and the lagged effect of broadband in any subscribers through lower prices and increased network effects economy are key variables to take into account in the model. (that is, consumer surplus). Figure D.3: Network externalities and critical mass effect Methodology Network Externalities and Critical The macroeconomic analysis can be broken down into Mass Effect four parts: • Part A. Using panel data econometric analysis, we determine the impact of price, network effects, and broadband A critical element of the evolving theoretical framework of network externalities of broadband is the impact that availability on broadband adoption in African countries in infrastructure penetration levels may have on output. Is there a general. This provides the parameters used to measure the linear relationship between broadband adoption and economic increase in broadband adoption resulting from the SDM. growth? Or are we in the presence of a more complex • Part B. We then estimate the impact of integration in causality effect? East Africa in two different scenarios: Base and High. We The ‘critical mass’ findings of research of the impact of estimate (a) how integration impacts price reductions for telecommunications on the economy indicate that the impact broadband, (b) the level of network effects across the six of broadband on economic growth may only become significant once the adoption of the platform achieves high penetration countries resulting from integration and creation of a larger levels. Theoretically, it appears that there is a nonlinear (or market, and (c) the increase in broadband availability for S-Shaped) relationship between broadband penetration the two different scenarios. Combining these estimates and output. with the parameters from Part A, we can determine the At low levels of broadband penetration, the impact of increase in broadband adoption in each of the six countries. broadband on the economy is minimal due to the ‘critical mass’ • Part C. Using the increase in broadband adoption figures concept. The impact on the economic output is maximized from Part B, we estimate impacts on GDP and jobs for once the infrastructure reaches a critical mass point, generally both the Base and High Scenarios. For jobs, previous ITU associated with levels of penetration of developed countries. Then, once it reaches the saturation point, impact on economic findings on broadband’s impact on jobs is utilized. In the output diminishes and tends to zero. case of GDP, an additional panel data econometric analysis The implication of this finding for developing countries is was conducted to determine broadband’s impact on GDP significant. Research points to the fact that to achieve an growth based on penetration level. important level of economic impact, broadband needs to reach • Part D. Finally, in addition to an increase in the number high levels of penetration. In this regard, it is worth highlighting of users, the SDM will deliver benefits to existing users, Koutroumpis (2009)185, who finds that the contribution of through access to a wider network of users, a broader range broadband to economic growth increases with penetration of digital content and services, as well as through lower (based on a study of the OECD countries). According to this research, in countries with low broadband penetration broadband prices. We estimate these consumer surplus (under 20 percent), an increase of 10 percent in broadband benefits for the Base and High Scenarios as well. adoption contributes to 0.08 percent to GDP growth, while in • To summarize the methodology, as an example, we look at countries with medium penetration (between 20 percent and how we would estimate the impact of changes in one of 30 percent), the effect is a 0.14 percent contribution, and in the variables—a price fall in each country resulting from the countries with penetration higher than 30 percent, the impact of 10 percent adoption reaches 0.23 percent. SDM—as set out in Table D.1. 185 See table D.2. for full reference. 61 Table D.1: Summary of methodology Part Objective Determine the percentage increase in broadband demand a caused by a 1 percent Part A decrease in the price of broadband. We will do the same for network effects and broadband availability. Estimate the percent b that broadband prices will fall, based on the SDM, to get the Part B total increase in broadband adoption. Multiplying by a gives the increase in broadband adoption expected in each country, c. Calculations for the Base and High Scenarios. Use ITU multiplier for jobs and the GDP multiplier calculated, for the purpose of this Part C paper, to determine GDP and job increases in each country, c. Calculations for the Base and High Scenarios. Show for each country how the decrease in price b will save money for existing users, and how the increased network effects from the SDM will make the Internet more Part D useful for these users, raising their consumer surplus. Calculations for the Base and High Scenarios. Part A: Elasticities of demand for broadband In this section, we estimate the impact of integration on to assess the elasticities of demand for broadband access (that is, broadband penetration using econometric analysis, based on how sensitive demand is to price, network effects, and availability existing data from the ITU, GSMA Intelligence, and the World of broadband), which will then be utilized to determine increase Bank. To ensure relevance, the analysis uses data from all in broadband penetration due to the SDM. The equation used to countries across Africa and uses a time series from 2012 to 2016 determine these elasticities is featured in Table D.2. to increase the significance of the results. The estimates are used 62 A SINGLE DIGITAL MARKET FOR EAST AFRICA Table D.2: Equation used to determine elasticity of demand Variable Definition Dependent variable: This variable measures the number of people with mobile broadband subscriptions. This Penetration rate for mobile broadband is for unique subscribers rather than the number of total connections, to control for some users having multiple subscriptions/SIMs. Independent variables: The variable measures the cost of 500 MB of mobile broadband data with a prepaid plan, Price of mobile prepaid service using a mobile handset, in U.S. dollars. This measures the amount of international capacity that is used for Internet services. As more users go online and use more services, they are generating more traffic, which uses more bandwidth. As such, bandwidth is a good metric to measure total Internet usage in a country. While there is no metric available for total bandwidth used within a country, international bandwidth used is a good estimate, as currently up to 90% of usage is international in countries in East Africa. Furthermore, the amount of bandwidth used is a good proxy for network effects, which arise when the benefits of a service for each user increases exponentially with the number of users of that service. By and large, traffic is generated by communications between users, International bandwidth in Mbit/s including emails and video conferences, and by interactions with online content and services, such as YouTube and websites. The former traffic represents direct network effects—the more users there are to communicate with, the more traffic and the greater the benefit of going online. The latter traffic represents indirect network effects—the more users there are, the greater the amount of data, content, and services are available, and the greater the benefit of going online. As a result, the amount of online usage, represented by international bandwidth used, is a good proxy of network effects. A market with higher online usage and corresponding traffic has more users to communicate with, and more content and services being used, making it more attractive for new users. This is measured by the percentage of the population that can receive at least a 3G mobile signal. This is a necessary, but not sufficient, precursor of Internet adoption in a country. As Availability of broadband economies of scale for deployment of connectivity infrastructure and services grow across the SDM and demand rises from the previously discussed effects, adoption will grow as network coverage increases. This measures the spending level in each country, which is an important enabler of adoption of broadband, all other factors being equal. While the exercise in the present paper shows Household consumption per capita in USD the benefits of the SDM on GDP and jobs, which will ultimately affect income consumption levels, we are holding the spending level constant for this analysis, to focus on the direct effects of the SDM through the previous three metrics. 63 Table D.3 summarizes the 2016 figures for these variables in each We summarize the results from the econometric regressions of the six countries. This is the starting point for our consideration in Table D.4, with further explanation of the methodology in of the economic impact of integration. Annex D.1. Because the variables are logged, the results can be interpreted as follows: for international bandwidth, the results Table D.3: Pre-integration data for East Af- indicate that a 100 percent increase in international bandwidth rican countries in any country is associated with a 14 percent increase in mobile penetration in that country. Likewise, for the other independent variables, each of which impacts mobile penetration based on the South Sudan parameter in Table D.4. Variable Tanzania Rwanda Burundi Uganda Table D.4: Parameters from regression results Kenya Log of Mobile Unique Results187 Penetration mobile broadband 51.71 29.13 23.84 23.84 8.54 10.95 subscribers (as a % of the Log of price −0.068 population) Log of Intl_ 0.15 Mobile Bandwidth broadband 2.46 2.30 2.54 5.84 9.31 4.83 price (US$) Log of consumption 2.20 per capita International internet Log of availability 0.46 860,570 12,190 19,024 49,984 352 3,510 bandwidth (Mbit/s) Part B: Impact of the SDM on prices, Mobile network effects, and broadband broadband availability availability (% 82 85 92 45 20 40 of population In this section, we make assumptions about how the SDM will covered by signal) affect prices, network effects, and availability of broadband in each of the six countries for two different scenarios: Base and Source: GSMA, ITU, and World Bank, 2016.186 High. We then use the assumptions about each of the three variables to see the impact of integration on mobile penetration 186 Unique mobile user penetration is from the GSMA. The data series used has subsequently been revised by GSMA after this study was completed; mobile across the countries, the key variable to determine the impact on broadband price is from the ITU; International internet bandwidth is from the GDP and jobs in Part C. ITU, household consumption per capita (not shown) is from the World Bank, and mobile broadband availability is from the GSMA, except for Tanzania, which is taken from the ITU as one operator does not report to the GSMA. 187 All coefficients are statistically significant - see Annex D.1 for details. 64 A SINGLE DIGITAL MARKET FOR EAST AFRICA • Price. Based on the removal of barriers across the six Table D.5: Post-integration variables for East countries, including the elimination of mobile data roaming Africa Base and High Scenarios rates and duties on handsets, we estimate that the prices in the Base Scenario will equalize across all six countries, Base South Sudan decreasing to US$2. This is a slightly lower rate than the Scenario Tanzania Rwanda Burundi current lowest market rate in the region (US$2.30), which is Uganda Kenya offered in Tanzania. For the High Scenario, we estimate that prices will decrease to US$1.5 across the six countries. In the long run, both scenarios may be conservative, as economies of scale and competition are likely to further reduce rates Mobile even in those three countries where prices are currently broadband 2 2 2 2 2 2 low. However, in the short run, the benefits in South Sudan, price (US$) Burundi, and Uganda, where prices are currently much higher, are most significant. • International bandwidth. In an integrated market, internet International Internet users in each country will benefit to a degree from the usage 881,835 36,570 57,072 149,952 1,056 10,530 bandwidth in neighboring countries, due to the elimination of barriers, (Mbit/s) allowing the free flow of data and services across borders. To again be conservative, in the Base Scenario, we assume that network effects rise in each country by just 25 percent of the Mobile total international bandwidth in the other five countries. For broadband 100 100 100 56 25 50 the High Scenario, we assume that network effects will rise in availability (%) each country by 50 percent. Given the significant amount of bandwidth used in Kenya before integration,188 the other five countries receive a much more significant benefit. 189 High South Sudan Scenario • Availability. We assume that the SDM will lower the costs of Tanzania Rwanda Burundi Uganda deployment while also increasing demand, thus resulting in Kenya greater mobile broadband availability than would otherwise be the case. We assume that the increase would be 25 percent for the Base Scenario and as much as 50 percent for the High Scenario. Please note that the availability was capped at 100 Mobile percent for Kenya, Rwanda, and Tanzania, which were already broadband 1.5 1.5 1.5 1.5 1.5 1.5 close to full population coverage. price (US$) Table D.5 summarizes the assumed values of the variables after integration for the Base and High Scenarios. International 188 Note that we tested the extent to which Kenya, along with South Africa, act as Internet 903,100 60,950 95,120 249,920 1,760 17,550 outliers, due both to the high usage in their countries and that the countries act bandwidth as hubs for international bandwidth coming in from undersea cables and going (Mbit/s) out to neighboring countries. As noted in the Annex, removing Kenya and South Africa from the regressions (which include all African countries) or subtracting potential hubbed traffic does not affect the results statistically, and thus, the results are robust to these outliers. Mobile broadband 189 We also assume that the network effects cannot more than triple the pre- 100 100 100 90 68 69 integration international Internet bandwidth, as particularly in the countries availability with the very low starting point, the total network effects were increasing by (%) magnitudes based on the large starting point of Kenya. We assume that there is a limit to the benefits that could be absorbed from such a low starting point. 65 Applying the regression parameters in Table D.4 to the new are achieved. From there, we expect that growth in broadband post-integration levels of price, international bandwidth, and penetration would continue to accelerate, as the changes begin availability gives the following mobile penetration levels in each to reverberate throughout the economy, government, and users country for each of the scenarios being studied. This is the impact (based on the increased availability and value of online services, after the integration has been implemented, and the resulting content, and commerce, plus the falling cost and increasing post-integration broadband prices, availability, and networks quality of broadband). Table D.6: Post-integration mobile broadband penetration for East Africa Base and High Scenarios South Base Scenario Kenya Tanzania Rwanda Uganda Burundi Sudan Mobile broadband penetration before 52 29 24 24 9 11 integration (%) Mobile broadband penetration after 58 37 30 34 12 15 integration (%) Change in percentage points (%) 6 8 6 10 3 4 South High Scenario Kenya Tanzania Rwanda Uganda Burundi Sudan Mobile broadband penetration before 52 29 24 24 9 11 integration (%) Mobile broadband penetration after 67 55 52 46 22 19 integration (%) Change in percentage points (%) 15 26 28 22 13 8 As described in the tables in this annex, all countries benefit from points in the Base Scenario to 8–28 percentage points in the integration with their neighbors, as a result of lower prices, greater High Scenario. Countries that started with the midrange mobile network effects, and broadband availability, with increases in penetration levels (that is, Tanzania, Rwanda, and Uganda) see mobile penetration due to SDM ranging from 3–10 percentage the greatest increase in penetration in both scenarios. 66 A SINGLE DIGITAL MARKET FOR EAST AFRICA Part C: Impact on GDP and jobs there is an increase in the number of jobs of between 0.2 percent and 0.4 percent.190 To be conservative, we used 0.3 percent in As noted in the literature review, increases in broadband the middle of the range. This leads to a job increase, detailed in adoption have a macroeconomic impact on GDP and jobs. To Table D.7, across the six countries—with an overall increase in determine the SDM’s impact on jobs, we use ITU studies, which jobs at 2 percent in the Base Scenario and 6.2 percent in the show that for every 1 percent increase in broadband penetration, High Scenario. Table D.7: Post-integration jobs forecasts for East Africa Base and High Scenarios Base Scenario Kenya Tanzania Rwanda Uganda South Burundi Sudan Jobs 2016 17,143,199 23,341,941 5,869,284 17,921,987 4,547,584 4,776,926 Post integration 17,456,220 23,918,836 5,972,057 18,441,591 4,630,436 4,837,709 Change 313,021 576,894 102,773 519,604 52,852 60,783 % Change 2 2 2 3 1 1 High Scenario Kenya Tanzania Rwanda Uganda South Burundi Sudan Jobs 2016 17,143,199 23,341,941 5,869,284 17,921,987 4,547,584 4,776,926 Post integration 17,906,047 25,143,348 6,369,997 19,102,131 4,756,095 4,897,171 Change 762,848 1,801,406 500,713 1,180,145 178,511 120,245 % Change 4 8 9 7 4 3190 190 See http://www2.itif.org/2013-tech-economy-memo.pdf citing ITU results. 67 For GDP, we have conducted an additional set of panel data As with jobs, these results will be realized when the post- econometric regressions, based on the work done by Edgardo integration changes are achieved, with the time horizon set Sepulveda for the World Bank in the paper titled ‘Broadband at five years for this study. From there, the GDP growth rate & Economic Development: Regression Analysis’ from 2017. would increase according to the rate cited earlier, and the further The paper studied broadband’s impact, based on income increase in broadband penetration and GDP will lead to further level, dividing a set of 110 countries into two groups: (a) low increases in jobs. and medium income and (b) high income. To address the well-documented endogeneity issue191 when studying GDP Note also that these results emphasize the positive impacts on and broadband, the paper used a two-stage IV estimation the countries currently lagging in online activities, suggesting method. Specifically, the econometric methodology presented that they will benefit greatly from access to a larger market, in Czernich (2011)192, which uses a nonlinear IV estimation resulting increase in mobile broadband penetration as well as method. For additional information on the model and its access to the online content and services already developed specifications, see Annex D.2. in the larger markets. On the other hand, the more advanced countries such as Kenya have less significant benefit in terms For this paper, the team used the same two-stage IV estimation of mobile broadband penetration as their markets are already method and took a step further with respect to the 2017 well-advanced. paper by dividing the two sets of countries (that is, the set of 110 countries from the original paper and the subset of Sub- However, the model and resulting analysis do not quantify the Saharan countries), depending on penetration level—(a) below significant benefits that existing digital companies will experience, 30 percent, (b) from 30 percent to 60 percent, and (c) above 60 given the removal of barriers to serving an expanded regional percent), instead of income level. As explained in Annex B, Sub- customer base beyond their national borders. As a result, for Saharan Africa regressions did not show sufficiently significant countries with a more advanced digital industry such as Kenya, results, due to the limited number of countries with complete the GDP and job benefits are likely to be higher than measured data sets for multiple years. The team did find significant results using this methodology. (at a 1 percent confidence level) for low and medium broadband As shown in Tables D.7 and D.8, the benefits of integration, penetration levels (30 percent to 60 percent penetration) using based on the increase in broadband adoption, are positive for the entire set of 110 countries. The results suggest that a 10 both scenarios and will increase over time, particularly in those percent increase in mobile broadband in a country with countries that are lagging in broadband penetration today. The medium or low penetration increases annual GDP per capita average increase in GDP for the Base Scenario is 0.57 percent and by 0.80 percent, slightly higher than the results obtained in the 1.6 percent for the High Scenario, while the average increase in 2017 paper. This leads to a GDP increase, detailed in Table D.8, employment is 2.2 percent for the Base Scenario and 6.2 percent across the six countries. In total, GDP growth due to SDM is for the High Scenario. Across the region, the increase in GDP estimated to reach US$0.93 billion, meaning a 0.57 percent totals US$0.93 billion and the increase in jobs totals 1.6 million increase in average for the Base Scenario, and US$2.6 billion, or for the Base Scenario and US$2.6 billion and 4.5 million for the 1.6 percent for the High Scenario. High Scenario. 191 Ibid. 192 See table D.2. for full reference. 68 A SINGLE DIGITAL MARKET FOR EAST AFRICA Table D.8: Post-integration GDP forecasts for East Africa Base and High Scenarios South Base Scenario Kenya Tanzania Rwanda Uganda Burundi Sudan GDP 2016 (US$, millions) 70,529 47,431 8,376 25,527 9,015 3,007 Change 343 312 39 197 27 10 (US$, millions) % Change 0.49 0.66 0.47 0.77 0.31 0.34 South High Scenario Kenya Tanzania Rwanda Uganda Burundi Sudan GDP 2016 (US$, millions) 70,529 47,431 8,376 25,527 9,015 3,007 Change 836 976 190 448 93 20 (US$, millions) % Change 1.19 2.06 2.27 1.76 1.04 0.67 69 Part D: Welfare impact for existing Figure D.4: Mobile broadband adoption demand broadband users before integration In the previous sections, we have examined the increase in mobile penetration due to integration, which brings new users online, additionally increasing GDP and jobs. However, the same factors which lead to growth in the number of mobile broadband users Mobile broadband price also benefit existing users. Existing users benefit from lower prices for the same (or improved) services and from increased network effects which spur the creation of a greater variety and higher quality of digital content and services, making internet usage more valuable to them. In this section, we seek to estimate the increase in these benefits, P in terms of the overall increase in consumer surplus, which Demand curve (D) includes some cost savings for the users. As price falls, consumers save on expenditures. Furthermore, larger network effects make Mobile broadband adoption (quantity) adoption more attractive for everyone, increasing demand, which Q is represented by shifting out the demand curve, raising their consumer surplus further, as highlighted in the following figures. Figure D.5: Mobile broadband adoption demand Figure D.4 shows the demand before integration. For each after integration country, at the particular price level, there is a demand for mobile broadband represented, based on the intersection of the price and the demand curve. In Figure D.5, there are two effects. First, the price falls, so that for all existing mobile broadband subscribers, there is an expenditure saving represented by the light blue rectangle. At the same time, the demand curve shifts Mobile broadband price out to reflect the increased value of being online, based on the increased direct and indirect network effects. For existing users, this means that their willingness to pay increases, as there is more value to being online, based on the P’’ larger user base in the region as well as the availability of new services that are likely to be made available as a result. This P increase is represented by the dark blue rectangle. The two P’ D’ rectangles together represent the increase in consumer surplus Demand curve (D) for existing users—the difference between what they would have Mobile broadband adoption (quantity) been willing to pay, based on the demand curve, and what they actually pay, represented by the lower price level.193 Q 193 We note that this increase in consumer surplus is measured conservatively, because the dark blue rectangle only represents the increase in willingness to pay for the marginal user. For all the other users, with a higher willingness to pay, there is additional increase in consumer surplus, measured as the area between the demand curves. However, this is difficult to estimate, and as shown, the increase in consumer surplus is already significant. 70 A SINGLE DIGITAL MARKET FOR EAST AFRICA Table D.9 provides the numbers for each country, both in terms higher pre-integration prices, as well as greater network effects. of the gain to individual users in consumer surplus and the Comparably, existing users in Kenya, with relatively low initial aggregate total for all pre-integration subscribers. We note that mobile broadband prices and the largest pre-integration network the price impact is most important in the countries with the effects (measured by international bandwidth) gain slightly less highest pre-integration prices and that the network effects also from integration. However, it is likely that the benefits in Kenya have an impact on consumer surplus as countries have access are underrepresented, as the methodology does not account for to the large network effects enjoyed by consumers in Kenya as the increase in new services and the improving quality of services regional barriers are removed. This is because of the significance available as a result of the larger regional market. In particular, of network effects in the regression results. over time, new innovative content and services will be made available to a much larger regional market. Their availability will be Existing users in Tanzania and Rwanda gain less because of their reflected by further increases in willingness to pay in all countries, relatively low prices pre-integration, while those in Uganda, including Kenya, resulting in further gains in consumer surplus. South Sudan, and Burundi enjoy significant gains because of their Table D.9: Post-integration increase in consumer surplus for East Africa Base and High Base Scenario Kenya Tanzania Rwanda Uganda South Sudan Burundi Consumer surplus gain per 0.60 24 27 64 103 53 sub US$ Total consumer surplus US$ 14,937,734 387,768,708 75,750,421 632,018,885 107,642,989 60,503,986 High Scenario Kenya Tanzania Rwanda Uganda South Sudan Burundi Consumer surplus gain per 1.24 79 87 202 323 167 sub US$ Total consumer surplus US$ 30,978,479 1,272,214,727 247,025,875 1,996,510,102 337,047,726 192,088,703 71 The gains in consumer surplus are significant in both scenarios due at the two ends of the income spectrum covered by the Brazilian to the price decreases and network effects in all countries except data. At the lower end, in Figure D.6, are households with income Kenya, which already had a significant network and relatively low less than the minimum wage in Brazil (<1SM ), while Figure D.7 prices. The countries with the highest prices before integration presents households with income levels greater than 10 times (Uganda and South Sudan) enjoyed the greatest increase in the minimum wage for Brazil (>10SM). The lowest income group consumer surplus per existing user. went from 3 percent to 29 percent internet adoption between 2010 and 2016, while the highest income group went from 86 Microeconomic impact percent to 97 percent over the same years. This section examines the impact of the SDM on broadband Figure D.6: Percentage of Brazilian households adoption and subsequent economic benefits from increasing in the <1SM segment with Internet access broadband penetration at the bottom of the pyramid. This review consists of two parts: 100% • An assessment of distributional impact—exploring how 80% the impacts of the SDM, particularly with respect to price decreases, will be distributed across lower income levels, as measured by broadband adoption levels. 60% • A brief case study of the SDM’s impact on access to a key digital service, mobile money—exploring the impacts across 40% the economies of East Africa. 20% Adoption levels for lower-income groups -% For this review, we take advantage of the detailed survey data 2010 2011 2012 2013 2014 2015 2016 collected in Brazil by the Regional Center for Studies on the Development of the Information Society. This survey has been conducted since 2005, collecting detailed demographic data Figure D.7: Percentage of households in the on households, including income levels, and information about >10SM segment with Internet access Internet adoption and usage.194 This data set is unique, particularly for an emerging economy and provides the best insights into the 100% questions posed in this study. Similar data are not available within the East Africa subregion or wider Sub-Saharan Africa. 80% While the data is from Brazil, which has a higher per capita income level than countries in East Africa, as well as greater 60% levels of fixed broadband, it is nevertheless broadly instructive in terms of the likely adoption of broadband in response to service 40% availability and pricing, particularly at the base of the pyramid. We focus on household adoption data (as opposed to individual 20% usage), as this data is more detailed in terms of family spending on broadband. We also use the years 2010–2016, because there -% is no ITU broadband price data available before 2010. 2010 2011 2012 2013 2014 2015 2016 Figures D.6 and D.7 show how overall internet adoption increased 194 For more details, see the 2016 Survey at https://www.cgi.br/media/docs/ Source: The Brazilian Internet Steering Committee, 2017 195 publicacoes/2/TIC_DOM_2016_LivroEletronico.pdf. Page 163 discusses 195 See https://www.cgi.br/media/docs/publicacoes/2/TIC_DOM_2016_ the methodology. LivroEletronico.pdf. 72 A SINGLE DIGITAL MARKET FOR EAST AFRICA Over the period covered, fixed broadband prices in Brazilian real Figures D.9 and D.10). The figure D.9 shows the divide among (R$) have stayed relatively constant, as can be seen in Figure households that have broadband, while the one below (Figure D.8. At the same time, the mobile broadband prices came down D.10) shows the divide as a percentage of all households in the relatively significantly in the first year that the ITU measured lowest income group and therefore demonstrates how mobile them and then stayed below the fixed broadband prices. While broadband adoption drove growth in household adoption. the prices overall look fairly stable, the dashed line in Figure D.8 shows the increase in GDP per capita in R$ over the same period, Figure D.9: Internet connection type used by with a significant increase. Thus, in terms of affordability against households with Internet access (<1SM) income, prices would fall relatively. Figure D.10: Internet access by connection Figure D.8: Fixed and mobile monthly 100% subscription charge (R$) on the left-hand axis Mobile broadband and GDP per capita on the right-hand axis (R$) 80% Dial up Fixed broadband 60% 80.00 40,000 Fixed broadband 40% Handset - Prepaid 500MB GDP per capita 60.00 30,000 20% -% 2010 2011 2012 2013 2014 2015 2016 40.00 20,000 type, as a % of total households in the <1SM income group 20.00 10,000 Source: ITU, World Bank, Analysys Mason, 2018. 35% Mobile access 30% - - Dial up 2010 2011 2012 2013 2014 2015 2016 25% Fixed broadband 20% Source: ITU, World Bank, Analysys Mason, 2018. 15% Likely in response to these price differences, households at 10% different income levels in Brazil expressed a significant difference in adoption levels of mobile versus fixed broadband. As more 5% users came online, much of the growth at the lowest income -% levels came from households taking advantage of mobile 2010 2011 2012 2013 2014 2015 2016 broadband offerings. As a result, by 2016, these households were split relatively evenly between fixed and mobile broadband access, while the final dial-up users shifted to broadband (see 73 This differed significantly from the highest income levels, whose This suggests that increased mobile broadband availability at growth in adoption came primarily from increased adoption of lower relative prices is a significant driver of broadband adoption fixed broadband access over the same period. In fact, mobile in Brazil for lower-income households, even when fixed broadband broadband usage in these households declined, both as a is available. In most of Africa, including East Africa, where mobile percentage of households in the income group who were online broadband is the predominant means of individual and household (see Figure D.11) and as a percentage of all households in the access, this implies that the households at the bottom of the income group online (see Figure D.12). pyramid benefit significantly in terms of affordable online access. Figure D.11: Internet connection type used by households with Internet access (>10SM) In terms of total spending on broadband, the trends again differ significantly at different ends of the income spectrum. Figure D.13 presents the monthly spend on broadband for the <1SM households. At this level, the least expensive broadband became relatively more popular, shifting up from 10 percent to 24 100% percent of households. By examining the spending patterns of the lowest income group (see Figure D.14), as an absolute number 80% of households (as opposed to within the group that has adopted broadband), we can see how, as Internet adoption grows over 60% Fixed broadband time, an increasing number of households take up the lower-cost 40% services. We can thus conclude that the lower-cost offerings are Dial up driving adoption of internet in the lowest part of the pyramid. Mobile access 20% Figure D.13: Monthly spend on broadband (<1SM) -% 2010 2011 2012 2013 2014 2015 2016 10SM income group 10SM income group 10% dummy Significance * N/S Interaction (MBB > 10% * 0.067 0.027 MBB rate) Significance N/S N/S Predicted MBB 0.043 0.032 > 35% dummy Significance N/S N/S Interaction (MBB > 35% * −0.129 0.075 MBB rate) Significance * N/S Predicted MBB 1.537 0.178 > 65% dummy Significance N/S N/S Interaction (MBB > 65% * −2.249 −0.264 MBB rate) Significance N/S N/S r2 0.04 0.03 0.09 0.10 0.07 0.05 0.16 0.13 0.02 0.03 0.06 0.05 0.07 0.04 f-value 11.12 4.53 15.44 10.16 18.95 7.18 31.31 12.97 1.23 1.16 18.17 7.37 6.56 2.78 years 11 11 11 11 12 12 12 12 10 10 11 11 11 11 observations 770 770 770 770 792 792 516 516 230 230 902 902 770 770 Note: significance: *** = 1% conf.; ** = 5% conf.; * = 10% conf.; otherwise not significant (N/S); MBB = Mobile broadband; SSA = Sub-Saharan Africa; LAC = Latin America and the Caribbean 84 A SINGLE DIGITAL MARKET FOR EAST AFRICA New 2018 Results This section presents the results of the new regression analysis 8 Sub-Saharan African countries that were excluded in the 2017 carried out for this project. It first presents the results based on regressions because they did not meet set mobile broadband our initial approach, after which the results based on the revised penetration thresholds. None of the “without macros” regressions approach are summarized. converge in the first stage of the 2S-IV process, which means that an IV cannot be calculated for the second stage. Regressions #300L–#305L address the nonlinear non-convergence by using Initial Sub-Saharan Africa-Centric a linear specification in the first stage. Of the regressions that Approach converge, the multiplier results are generally negative and sometimes significant. The initial approach was based on the observation that the default multiplier of 0.056 was based on a “Global” sample (LM- To see whether the above-noted results were robust to specific 70) that included low- and middle-income countries from around Sub-Saharan Africa samples, regressions #450–#455 are based the world. There are 22 Sub-Saharan African (referred to as ‘SSA- on slightly different Sub-Saharan Africa samples that more 22’) countries in the LM-70 sample, or about 31 percent. In this closely match GDP and mobile broadband penetration of the EAC context, our initial approach was to attempt to calculate a Sub- countries. The results were similar to those reported earlier. Saharan Africa-specific multiplier using only Sub-Saharan Africa Regressions #500–#506 try a different variation, this time using data in the sample. the first stage results from LM-70 and applying the predicted Based on our results from the 2017 study, however, we IV only for a subsample of Sub-Saharan Africa countries. The recognized that the initial approach would not necessarily result “without macros” results are somewhat encouraging in that they in conclusive or reasonable results. First, we knew from #363 are positive, but they are not significant and have relatively low that Sub-Saharan Africa countries had a negative and significant parameter values. dummy variable. Further, many Sub-Saharan Africa countries are At this point, we were asking whether the results were in the lowest income group #59 and other similar regressions show reflecting Sub-Saharan Africa-specific factors or more general that low-income regressions generally do not provide significant global phenomenon related to countries with relatively lower results (which is why the low- and middle-income groups were penetration or later introduction. To test this hypothesis, we combined into LM-70 in the 2017 study. The results presented in started to expand the analysis and created a new global sample, the following paragraphs confirm these concerns; a Sub-Saharan LM-50 (a subsample of LM-70 that excludes 20 early adopters Africa-specific sample applying a 2S-IV methodology did not and high penetration countries). Regressions #310 and #314 are provide reasonable or significant results. Our revised approach based on LM-50 and have a positive (but insignificant) multipliers is presented in the following section. in the range of 0.019–0.035. Table D.2.2 presents the first set of regressions using our initial A different, quasi Sub-Saharan Africa-centric approach is to focus approach, and includes work based on the database compiled for on the Sub-Saharan Africa-dummy and interaction terms in the the 2017 study, which included data from 2015 (this is referred to context of a global sample. Regressions #320 and #330 (for LM- as “2017” in the data row of the following tables), as well as the 70 and LM-50 for robustness) in Table D.2.4 shows a significant database updated to 2016 for the current project (referred to as Sub-Saharan Africa interaction of about 0.043 to 0.045. This is “2018” in the data column). incremental to the corresponding base result, which is positive Simple pooled regressions #10 and #11 show a negative but and significant for #320, therefore the Sub-Saharan Africa-specific insignificant multiplier. Regressions #300-#305 and #400–#415 result multiplier would be the sum of these at 0.090. A more highlight the problem with the initial approach used, presented conservative approach would take the base as zero (because it is earlier. These include regressions for the original SSA-22 group, as not significant in the robustness check of #330) and suggest that well as the expanded SSA-30 group of countries that included the a more conservative Sub-Saharan Africa-specific multiplier is in the range of 0.043–0.045. 85 Table D.2.2: New 2018 regressions Type of Pooled OLS 2-stage IV (nonlinear) 2-stage IV (linear) 2-stage IV (nonlinear) Regression Regression # 10 11 300 301 304 305 400 401 414 415 300L 301L 304L 305L 310 311 314 315 Data 2017 2018 2017 Group SSA-22 SSA-30 SSA-22 SSA-30 SSA-22 SSA-30 LM-50 Countries SSA Global Period 2005–2015 2007–2015 MBB −0.014 −0.006 No −0.016 No 0.026 No −0.024 No 0.058 −0.090 −0.084 −0.059 −0.048 0.035 −0.050 0.019 −0.035 (predicted) Significance N/S N/S CNV N/S CNV N/S CNV N/S CNV N/S *** ** N/S N/S N/S N/S N/S N/S Macro No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes Controls? Observations 242 242 242 242 330 330 237 237 314 314 242 242 330 330 550 550 450 450 Note: Significance: *** = 1% conf.; ** = 5% conf.; * = 10% conf.; otherwise not significant (N/S); MBB = Mobile broadband; SSA = Sub-Saharan Africa. 86 A SINGLE DIGITAL MARKET FOR EAST AFRICA Table D.2.3: New 2018 regressions Type of 2-stage IV (non-linear) Regression Overall 500 501 502 504 506 regression 1st stage 123 124 123 123 123 regression Data 2017 Group LM-70 Countries Global Period 2005–2015 Observations 770 Regression # 450 451 452 453 450 454 455 321 300 301 300 300 300 Data 2018 2017 Group SSA-18 SSA-21 SSA-18 SSA-29 LM-70 SSA-22 Countries SSA Global SSA Period 2008–2015 2010–2015 2008–2015 2005–2015 2005–2015 2009–2015 2010–2015 2011–2015 MBB —0.192 — — — 0.064 — — —0.001 0.017 —0.003 0.007 0.020 0.012 (predicted) Significance * * N/S N/S N/S N/S N/S N/S N/S N/S N/S N/S N/S Macro No Yes No Yes No No Yes Yes No Yes No No No Controls? Predicted —-0.026 MBB > 10% Dummy Significance N/S Inter. (MBB —0.026 > 10% * MBB rate) Significance N/S SSA Country -0.022 Dummy Significance *** Inter. (SSA * 0.039 MBB rate) Significance * Observations 144 130 108 108 144 232 202 770 242 242 154 154 132 132 110 Note: Significance: *** = 1% conf.; ** = 5% conf.; * = 10% conf.; otherwise not significant (N/S); MBB = Mobile broadband; SSA = Sub-Saharan Africa. 87 Revised Global Approach The revised approach reverts to the global sample methodology Regressions #850–#852 in Table D.2.4 confirm the earlier Sub- while applying a project-specific country classification criteria to Saharan Africa-specific results: they show that even under the rank countries and populate the country categories. This revised most favorable assumptions, that general Sub-Saharan Africa- approach also starts to explore whether the multiplier varies by the specific multipliers are relatively low—that is, in the range of level of mobile broadband penetration. To explore how a project- 0.030—and not significant. specific country classification would affect the regression results, we created the following ranking criteria based on a weighted average of Using the new classification criteria, Table D.2.6 examines whether 2013–2015 mobile broadband penetration (“MBBpen”) of the 110 the multiplier varies by level mobile broadband penetration. The countries in the 2017 study data set (based on 2, 3, and 4 weighting): focus of analysis should be on results from LMB-73 because it has a 30 percent Sub-Saharan Africa weight (compared to MB-39 with Weighted Average Mobile BB Penetration = ([2013 MBBpen * about 8 percent). In this regard, what regressions #900a–#905 do 2] + [2014 MBBpen * 3] + [2015 MBBpen * 4])/9 is split the sample by MBBpen ranges and looks at dummies and interaction terms. In this context, the LB-34 group are the 34 countries with lowest rank, which coincides with weighted mobile broadband penetration The dummies and interactions are equivalent to running a series of 0–30 percent, while MB-39 are the next 39 countries (coinciding of subsample regressions only for those observations that meet with 30–60 percent). LMB-73 combines LB-34 and MB-39. The 37 the range: for example, #900a includes all observations for all high penetration countries are above 60 percent. 73 countries that are between 0 and 4.99 percent. That would include observations both from a Poland (early days) and Burundi Regressions #800–#805 in Table D.2.5 show that the new (more recent). Regression #900a (backed by #910a) shows a very classification does have an impact on the regression results. strong negative effect for 0–5 percent MBBpen. This is a large part Relative to the core regression #323 with 0.056, regression #804 of the sample 368 of 803, or about 45 percent. There is then a big has a multiplier that is significant and positive at 0.126. Note that jump in regression #901 to about (0.131 + 0.116) 0.247 (backed as in prior equivalent specifications LB-34 is not significant, but MB- by 0.258 from #911) for 5–25 percent, and regression #905 results 39 is positive and significant. in a value of (0.280–0.203) 0.077 (backed up by Regression #905). 88 A SINGLE DIGITAL MARKET FOR EAST AFRICA Table D.2.4: New 2018 regressions Type of Regression 2-stage IV (nonlinear) Overall 600 602 604 850 851 852 regression 1st stage regression Data 2017 Group LM-70 LMB-73 Countries Global Period 2005–2015 2005–2015 Observations 770 803 Regression # 320 322 324 330 332 334 320 322 324 300 Data 2017 Group LM-70 LM-50 LM-70 LM-50 SSA-22 Countries Global SSA Period 2005–2015 2007–2015 2009–2015 2010–15 2005-15 2009-15 2010-15 Mobile BB 0.047 0.058 0.053 0.003 0.022 0.010 0.002 0.005 0.002 0.038 0.031 0.011 (predicted) Significance * *** *** N/S N/S N/S N/S N/S N/S N/S N/S N/S Macro No controls? SSA country —0.020 —0.015 —0.020 —0.015 —0.008 —0.009 dummy Significance *** *** *** *** * *** Inter. (SSA * 0.043 —0.0167 0.045 —0.025 —0.003 —0.028 MBB rate) Significance *** N/S * N/S N/S * Observations 770 450 490 Note: Significance: *** = 1% conf.; ** = 5% conf.; * = 10% conf.; otherwise not significant (N/S); MBB = Mobile broadband; SSA = Sub-Saharan Africa. 89 Table D.2.5: New 2018 regressions (based on new classification: ranking of average weighted 2013–2015 MBBpen Type of Regression 2-stage IV (nonlinear) Regression # 804 805 802 803 800 801 Data 2017 Group LMB-73 MB-39 LB-34 Countries Global Period 2005–2015 2007–2015 Mobile BB 0.126 0.119 0.110 0.114 0.001 −0.008 (predicted) Significance *** *** *** *** N/S N/S Macro controls? No SSA country −0.019 −0.010 −0.015 dummy Significance *** N/S *** Inter. (SSA * MBB 0.034 −0.003 0.068 rate) Significance N/S N/S N/S Observations 803 429 306 Note: Significance: *** = 1% conf.; ** = 5% conf.; * = 10% conf.; otherwise not significant (N/S); MBB = Mobile broadband; SSA = Sub-Saharan Africa. 90 A SINGLE DIGITAL MARKET FOR EAST AFRICA Table D.2.6: New 2018 regressions (based on new classification: ranking of average weighted 2013–2015 MBBpen; explore different MBB for LMB-73 and MB-39 [nothing significant for LB-34]) Type of Regression 2-stage IV (nonlinear) Regression # 900a 901 902 903 904a 905 910a 911 912 913 914a 915 Data 2017 Group LMB-73 MB-39 Countries Global Period 2005–2015 MBB (predicted) 0.076 0.131 0.118 0.175 0.060 0.280 —0.004 0.033 0.108 0.167 —0.032 0.194 Significance *** *** *** *** N/S *** N/S N/S *** *** N/S *** Macro Controls? No Predicted MBB 0-4.9% Dummy 0.013 —0.010 0.077 —0.013 Significance ** N/S *** N/S Interaction MBB 0-4.9% * MBB rate) —0.603 —0.395 —1.282 —0.209 Significance *** * *** N/S Predicted MBB 5-24.9% Dummy —0.014 —0.029 —0.050 —0.073 Significance *** N/S *** N/S Interaction MBB 5-24.9% * MBB 0.116 0.175 0.258 0.287 rate) Significance ** N/S *** *** Predicted MBB 25-44.9% Dummy 0.017 0.002 —0.006 —0.048 Significance N/S N/S N/S N/S Interaction MBB 25-44.9% * MBB —0.033 0.024 0.031 0.100 rate) Significance N/S N/S N/S N/S Predicted MBB 45+% Dummy 0.023 0.082 Significance N/S * Interaction MBB 45+% * MBB rate) —0.103 —0.180 Significance N/S *** Predicted MBB 25+% Dummy 0.051 0.026 ignificance *** N/S Interaction MBB 25+% * MBB rate) —0.203 —0.096 Significance *** N/S Count 368 277 139 19 158 Observations 803 429 Note: Significance: *** = 1% conf.; ** = 5% conf.; * = 10% conf.; otherwise not significant (N/S); MBB = Mobile broadband. 91 Conclusion The literature review included in the 2017 study, presented in this annex, indicated that broadband generally had a positive impact on GDP growth and that its impact varied between 0.023 and 0.150. The 0.023 and 0.150 were relative outliers and the bulk of the results were clustered in the 0.040 to 0.120 range. This is what the literature would consider an academic ‘safe harbor’, whose results would generally be considered as reasonable. The base specification for low- and middle-income countries in the 2017 study (0.056), based on regression #323, was firmly in that range and reflected a conservative stance. In this context, the new result of 0.126 from regression #804 is just outside the ‘safe harbor’ range and could be defensible. However, it is the highest of a number of our results and one in which we arranged the data (country classification criteria) specifically for our study. Further the results generated using our initial approach suggest that a conservative approach is appropriate for application of global results to a specific Sub- Saharan Africa sample. In this context, the multiplier to be applied to all countries will be an unweighted average of the four above-noted results (0.056, 0.064, 0.076, 0.126) this is 0.080. 92 A SINGLE DIGITAL MARKET FOR EAST AFRICA Annex D.3: Sample Countries for Part C High Upper-Middle Upper-Middle Low 1 Australia Albania Armenia Benin* 2 Austria Algeria Bangladesh* Burkina Faso* 3 Bahrain Angola Bolivia Burundi* 4 Belgium Argentina Cambodia Central African Republic* 5 Canada Azerbaijan Cote d’Ivoire Chad* 6 Chile Belarus Egypt, Arab Rep. Congo, Dem. Rep. 7 Croatia Belize El Salvador Gambia, The 8 Cyprus Bosnia and Herzegovina Ghana Madagascar* 9 Czech Republic Botswana Honduras Malawi 10 Denmark Brazil India Mali 11 Estonia Bulgaria Indonesia Mozambique 12 Finland China Kenya Niger* 13 France Colombia Kyrgyz Republic Senegal 14 Germany Costa Rica Lao PDR Tanzania 15 Greece Ecuador Mauritania Togo 16 Hungary Fiji Moldova Uganda 17 Iceland Gabon* Mongolia Zimbabwe 18 Ireland Georgia Morocco 19 Italy Guyana* Nigeria 20 Japan Iran Pakistan* 21 Korea, Rep. Jamaica Philippines 22 Latvia Jordan Sri Lanka 23 Lithuania Kazakhstan Sudan 24 Luxembourg Macedonia, FYR Swaziland 25 Malta Malaysia Tunisia 26 Netherlands Mauritius Ukraine 27 New Zealand Mexico Uzbekistan 28 Norway Namibia 29 Oman Panama 30 Poland Peru 31 Portugal Romania 32 Slovak Republic Russian Federation 33 Slovenia South Africa 34 Spain Suriname 35 Sweden Thailand 36 Switzerland Turkey 37 Trinidad and Tobago Venezuela, RB 38 United Kingdom 39 United States 40 Uruguay Note: *Countries not included in mobile broadband regressions. 93 A SINGLE DIGITAL MARKET FOR EAST AFRICA Pr s ntin vision, str t ic fr m work, impl m nt tion ro dm p nd imp ct ss ssm nt Copyright Statement This work is available under the Creative Commons Attribution Non-Commercial 3.0 IGO license (CC BY NC 3.0 IGO) http://creative-commons.org/licenses/by-nc/3.0/igo. Under the Creative Commons Attribution Non- Commercial license, you are free to copy, distribute, transmit, and adapt this work for non-commercial purposes, under the following conditions: Attribution—Please cite the work as follows: World Bank. 2018. A Single Digital Market for East Africa - Presenting a joint vision, strategic framework, roadmap, economic impact and readiness assessment ©World Bank License - Creative Commons Attribution Non-Commercial CC 3.0 IGO Noncommercial - You may not use this work for commercial purposes. Translations - If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations - If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Responsibility for the views and opinions expressed in the adaptation rests solely with the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content - The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to the Publishing and Knowledge Division, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. First edition. May 2018. SDM EAST AFRICA