Alternative Data Transforming SME Finance May 2017 © INTERNATIONAL FINANCE CORPORATION 2017. All rights reserved. 2121 Pennsylvania Avenue, N.W. Washington, D.C. 20433 www.ifc.org The material in this work is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. IFC encourages dissemination of its work and will normally grant permission to repro- duce portions of the work promptly, and when the reproduction is for educational and non-commercial purposes, without a fee, subject to such attributions and notices as we may reasonably require. IFC does not guarantee the accuracy, reliability, or completeness of the content included in this work, or for the conclusions or judgments described herein, and accepts no responsibility or liability for any omissions or errors (including, without limitation, typographical errors, and technical errors) in the content whatsoever or for reliance thereon. The boundaries, colors, denominations, and other infor- mation shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Di- rectors of The World Bank or the governments they represent. The contents of this work are intended for general informational purposes only and are not intended to constitute legal, securities, or investment advice, an opin- ion regarding the appropriateness of any investment, or a solicitation of any type. IFC or its affiliates may have an investment in, provide other advice or services to, or otherwise have a financial interest in, certain of the companies and parties (including those named herein). All other queries on rights and licenses, including subsidiary rights, should be addressed to IFC’s Corporate Relations Department, 2121 Pennsylvania Avenue, N.W., Washington, D.C. 20433. International Finance Corporation is an international organization established by Articles of Agreement among its member countries, and a member of the World Bank Group. All names, logos, and trademarks are the property of IFC and you may not use any of such materials for any purpose without the express written consent of IFC. Additionally, “International Finance Corporation” and “IFC” are registered trademarks of IFC and are protected under international law. Cover photo: www.istock.com G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION i Table of Contents Acronyms....................................................................................................................................................................................... iii Acknowledgements..................................................................................................................................................................... iv Foreword ........................................................................................................................................................................................ v Executive Summary..................................................................................................................................................................... vii Large SME Market, Credit Gaps, and Opportunities............................................................................................................... 1 The Rise of Digital SME Lending................................................................................................................................................. 5 4.1 SME Marketplace Lenders......................................................................................................................................................... 8 4.2 Technology, E-commerce, and Payment Giants................................................................................................................15 4.3 Supply Chain Finance Platforms............................................................................................................................................ 18 4.4 Mobile Data-Based Lending Models..................................................................................................................................... 20 Digital Lenders and Bank Convergence: The Future of SME Finance............................................................................... 23 5.1 Advantages and Challenges for Digital SME Lenders........................................................................................................ 23 5.2 Advantages and Challenges for Banks................................................................................................................................. 23 5.3 Collaborative Partnerships Gaining Momentum................................................................................................................ 24 5.4 The Rise of the Digital SME Banker/Lender......................................................................................................................... 28 Alternative Data Policy Issues and Challenges...................................................................................................................... 31 6.1 Data Privacy and Consumer Protection Issues...................................................................................................................31 6.2 Opt-in Models as Opposed to Opt-out Models.................................................................................................................31 6.3 Credit Reporting Service Providers....................................................................................................................................... 32 6.4 Cyber Security........................................................................................................................................................................... 32 6.5 Pricing Transparency................................................................................................................................................................ 33 6.6 Balancing Integrity, Innovation, and a Competitive Marketplace.................................................................................. 33 Conclusions and Recommendations....................................................................................................................................... 35 Endnotes.......................................................................................................................................................................................39 Annex 1 - Alternative SME Lender Business Profiles.............................................................................................................45 G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION iii Acronyms ACH Automated Clearing House MYOB Mind Your Own Business API Automated Programming Interface O2O Offline-to-online APR Annual Percentage Rate OECD Organisation for Economic B2B Business-to-business Co-operation and Development B2C Business-to-consumer OSHA Occupational Safety and Health B2G Business-to-government Administration CBA Commercial Bank of Africa P2B People-to-business CDR Call detail records P2P Peer-to-peer CRM Customer relationship management PAFI Payment Aspects of Financial CRF China Rapid Finance Inclusion EDC Electronic data capture POS Point-of-sale EFL Entrepreneurial Finance Lab ROA Returns on assets EFTPOS Electronic funds transfer at point of sale ROE Returns on equity FICO Fair, Isaac and Company RBS Royal Bank of Scotland FINRA Financial Industry Regulatory Authority RMB Renminbi GIS Geographic Information System SBI State Bank of India GPFI Global Partnership for Financial Inclusion SCF Supply chain finance GPS Global Positioning System SME Small and Medium Enterprise IFC International Finance Corporation SMS Short message service ICCR International Committee on Credit UK United Kingdom Reporting UPI Unified Payment Interface KCB Kenya Commercial Bank UPS United Parcel Service KYC Know Your Customer US United States MSME Micro, Small and Medium Enterprises VAT Value-added tax MVNO Mobile Virtual Network Operator WEF World Economic Forum iv ALTERNATIVE DATA TRANSFORMING SME FINANCE ACKNOWLEDGEMENTS This report was co-written by John Owens and Lisa Wilhelm. World Bank Group, Finance and Markets Global Practice: It has benefited from the peer review of many experts. The authors, the Project Leads on behalf of the SME Finance Oscar Maddedu Forum, Matthew Gamser, CEO, and World Bank Group, Fabrizio Fraboni Ghada Teima, Lead Financial Sector Specialist, would Alban Pruthi like to thank the following individuals for their valuable Fredesvinda Fatima Montes contributions. Ghada Teima Tony Lythgoe Amy Millington, eBay Foundation Douglas Pearce Thomas Deluca, AMP Credit Technologies Antonio Desocio and Banca d’Italia SME Finance Forum: Andreas Kind, IBM Allison Baller, IBM Matthew Gamser David Snyder, Wells Fargo Bank Hourn Thy Ross Leckow, IMF Nadia Afrin Michael Turner, Patrick Walker, Chet Wiermanski, PERC Jeffrey Anderson Dawei Liu, Creditease Daniel Drummer, JP Morgan Finally, the authors, the SME Finance Forum and the World David Medine and CGAP Bank Group would like to thank Germany, the Silicon Valley Enrico Libbiani, Experian Community Foundation, and the Swiss State Secretariat for Marco Benvenuto, Experian Economic Affairs (SECO) for their financial support for this Ian Milne, Experian work. We are grateful for the guidance and support from Luigi Mingotti, CRIF the co-chairs of the SME finance subgroup of the GPFI, the Dr. Tiandu Wang, People’s Bank of China German Federal Ministry for Economic Development and Leora Klapper, World Bank Cooperation (BMZ) and the Undersecretariat of Treasury, Benoit Fauvelet, la Banque de France Turkey. G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION v FOREWORD Access to financing remains one of the most significant Every time SMEs and their customers use cloud-based constraints for the survival, growth, and productivity of services, conduct banking transactions, make or accept dig- micro, small, and medium enterprises (SMEs). The G20 ital payments, browse the internet, use their mobile phones, countries, in developing their Global Partnership for Financial engage in social media, buy or sell electronically, ship pack- Inclusion, made SME finance one of its core workstreams. ages, or manage their receivables, payables, and record- Realizing the benefits to employment, innovation and the keeping online, they create digital footprints. This real-time, provision of many key goods and services depends on includ- and verified data can be mined to determine both capacity ing SMEs, not only individuals, in countries’ financial inclusion and willingness to repay loans. strategies. A rapidly growing crop of technology-focused SME lenders Digital SME finance, using alternative data, offers an extraor- are putting the use of SME digital data, customer needs, and dinary opportunity for addressing both sides of this problem. advanced analytics at the center of their business models, Growing digital finance will call on all the G20 High Level setting forth new blueprints for disrupting the SME lend- Principles for Digital Financial Inclusion. ing status quo. This report takes stock of the range of data, and the range of institutions using the data. It considers the The world’s stock of digital data will double every two years opportunities alternative data presents to narrow the financ- through 2020, fueled by the phenomenal intersection of and ing gap for SMEs. It also notes the new issues and poten- growth in mobile, cloud, big data, electronic payments, and tial risks raised by this massive increase and diversification social. By 2020, 60 percent of this digital data will come from of data supply to financial sector stability, and to consumer developing economies. Analytic and processing capabilities protection. are making great leaps, dispersing data-driven intelligence faster across these new digital ecosystems at plummeting This report was undertaken for the GPFI by the SME Finance transaction costs. Smart mobile devices are making this Forum and World Bank Group, with support from German information, computing power, and intelligence accessible Government, the Silicon Valley Community Foundation, and to SMEs and their financiers around the world. the Swiss State Secretariat for Economic Affairs (SECO). Natascha Beinker German Co-Chair, Global Partnership for Financial Inclusion German Federal Ministry for Economic Development and Cooperation (BMZ) G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION vii Executive Summary Access to financing remains one of the most significant constraints to the survival, growth, and productivity of micro, small, and medium REPORT SOURCES AND enterprises (SMEs). FOCAL POINTS The SME credit gap has proven to be an enduring structural feature across The report draws on ongoing prima- both developing and developed markets, even in countries that have en- ry and extensive secondary research acted a variety of policy measures to support SMEs and enhance financial covering: inclusion more broadly. In the world’s developing markets, about half of the estimated 400 million SMEs,1 or 180 to 220 million SMEs, still have • 800+ innovative digital SME lenders, unmet credit needs totaling US$2.1 to US$2.6 trillion.2 and digital commerce, payments, and service providers in more than The credit gap results from both demand and supply side problems. Many 60 countries across developed and SMEs are reluctant to seek or cannot access credit due to: the reams of developing markets (many emerging financial documentation and collateral requirements for obtaining a loan; only within the last five years); high costs and interest rates; and multi-week decision timeframes. • Recent alternative SME financing developments and trends; and Many banks consider SMEs to be high-risk clients, as well as high-cost • Ongoing discussions with digital SME clients to acquire, underwrite, and serve. Revenues per client are lower lender principals and industry leaders. relative to larger non-SME corporate clients. SME information is also often opaque. Therefore, many banks limit most of their lending to the largest The report focuses exclusively on the of the small firms. In the wake of the global financial crisis, increased cap- opportunities in digital SME data under ital and liquidity requirements, new regulations including those posed by -pinning SME lending that is largely un- Basel III,3 and shrinking returns on equity have made banks’ SME lending secured or secured by assets other than challenges even more daunting. Compounding these challenges are the real estate — and the kinds of SME dig- limited SME coverage by credit reporting service providers, weak contract ital lending platforms which are using or bankruptcy laws and judiciaries, and high SME informality in develop- it. (Digital SME data does, however, in- ing markets. clude consumer data highly relevant to assessing the risk of micro and small Digitizing SME finance and making use of transactional and alternative firm SME owners.) data offer an opportunity for addressing both sides of this problem. There- fore, this report focuses broadly on digital data for SME lending which The report does not address issues of includes new uses of both traditional data (bank, accounting, transac- equity, reward, donation crowdfunding, tional, and sales data) as well as alternative data (online ranking and social or consumer and real estate alternative media, mobile, and individual data, such as psychometric testing). financing platforms — except in cas- es where a particular lending platform According to the International Data Corporation (IDC), the world’s stock offers any of these types of financing in of digital data will double every two years through the year 2020. It will be addition to SME lending. fueled by the phenomenal intersection of and growth in mobile, cloud, electronic payments, and social media. Importantly, the IDC also says, The term “SME” as used throughout by 2020, 60 percent of this digital data will come from developing this report includes micro, small, and economies.4 Analytic and processing capabilities are making similar leaps, medium enterprises. (See Endnote 1 for dispersing data-driven intelligence faster across these new digital eco- SME definitions and other SME market systems — and at plummeting transaction costs.5 Smart mobile devices information, data sources, and estimation are making this information, computing power, and intelligence accessi- methodologies presented in this report.) ble to SMEs and their customers around the world. viii ALTERNATIVE DATA TRANSFORMING SME FINANCE Each time SMEs and their customers use cloud-based ser- Key findings include: vices, conduct banking transactions, make or accept digital payments, browse the Internet, use their mobile phones, • Banks have valuable data, but are often not using it: engage in social media, get rated online, buy or sell elec- Banks have a highly valuable repository of SME data, in- tronically, ship packages, or manage their receivables, paya- cluding SME owners’ customers’ daily transaction data bles, and recordkeeping online, they create and deepen the that provides reliable real-time visibility into SME cash digital footprints they leave behind. SMEs’ own, real-time, flows and credit capacity. However, most banks lack the and verified data — unprecedented volume, variety, and ability to create innovative SME lending models from it. velocity — also means more data can be used for credit The data often resides in a patchwork of legacy systems decision purposes. and data silos that make it difficult and costly to access. This gap has created an opening for digital SME lenders to A rapidly growing group of technology-focused SME lend- capture this market segment. ers are putting the use of SME digital data, customer needs, and advanced analytics at the center of their business mod- • Digital SME lenders are developing new relationships els, thereby setting forth new blueprints for disrupting the with SME customers and their data: In some cases, non- SME lending status quo. They can also offer more transpar- bank digital SME lenders insert themselves between banks ent, faster, easier, and better-tailored financing solutions and their SME customers, and forge fundamental chang- that today’s increasingly tech-savvy SMEs seek. es in SME customer expectations. SMEs are embracing the digital world more and more every day. Increasingly, SMEs are often willing to share their data in exchange for many SMEs are more tech-savvy, more sensitive to slower value, that is, access to credit and other value-added tools service and paper-intensive loan applications, and more that help them grow and become more productive. No willing to shop around for unmet and unserved financing longer are SME lenders limited to just underwriting a bor- needs. rower based on dated, often incomplete financial state- ments, missing or limited credit bureau information, or • New SME digital data streams are becoming more readily collateral that substitutes for a deeper understanding. available and accessible: Digital SME lenders leverage vast Instead, they are now able to develop a more comprehen- and expanding stores of data, including from electronical- sive view of the borrower’s business — one that illuminates ly verifiable, real-time sales, bank account money flows previously invisible SME strengths and weaknesses. and balances, payments, social media, trading, logistics, business accounting, and credit reporting service provid- The basis for this amplified view is a real-time flow of the ers, as well as a wide range of other private and public data SME’s digital footprint from the borrower to the lender that sources used in the SME credit assessment process. creates and continuously updates a rich model of the busi- ness.6 Digitally native lenders leverage the advances in com- • There are a wide range of digital SME originator lending puting power to match smart algorithms to these massive business models: The new digital SME lending originator data streams, including banking transaction information and business models that take advantage of the expanding money flows— and at increasingly lower costs. Data is gath- universe of SME digital data vary widely. This report high- ered in a variety of ways, including through: partnerships lights these business models, selected players, and the between Fintech providers and banks; Fintechs with direct digital SME data they use. It includes marketplace lenders, access to transactional data via bank Automated Program- tech, e-commerce, and payment giants which are extend- ming Interfaces (APIs); or the use of screen scraping tech- ing SME lending into their non-banking digital ecosystems nology.7 The more diverse the data and the faster the data where they are already dominant. It also includes supply can be analyzed, the more predictive its value will be. chain financing firms, mobile micro-lenders graduating to SME lending, and innovative banks. This report identifies the landscape of alternative data be- ing increasingly used to expand access to SME finance. It • Digital SME lending is becoming more of a global trend: also explores some of the new operating models and the That these innovators are sometimes simultaneously new breed of SME digital lending originators. It also looks launching nearly identical products in developed and de- at collaborative partnerships that are harnessing alternative veloping markets alike demonstrates just how profoundly and transactional data. and the report concludes by listing alternative data and technology are leveling the playing potential areas that policymakers and regulators need to field. As such, they are enabling new digital SME lenders in understand in order to enhance SME access to finance. many parts in the world to leapfrog traditional bank SME financing barriers.8 G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION ix • Digital SME lender-bank collaboration is also a growing These new entrants bring new complexities, risks, and ways part of the future of SME finance: Banks may have been of thinking about the SME financing value chains, as well as blind to digital SME lenders at first, and digital SME lenders new agenda items for policymakers and regulators. may have said they would replace banks. However, both parties now have come to a simple conclusion: there are This report identifies some of the key digital data agenda limits to what each player can do on their own and there items emerging as these innovations gain market traction to is strength in collaborating. Apart from partnerships with advance the stakeholder dialogue. The various policy theme banks, some non-bank digital SME lenders are instead areas that have surfaced for stakeholder consideration and partnering with each other, tech giants, cloud-based SME further discussion, include: service providers, or alternative lenders in other sectors. In • Data privacy and consumer protection issues other cases, they are securing their own banking licens- • Opt-in as opposed to opt-out models es, suggesting some new non-bank digital SME lenders • Credit information sharing still plan to forge an alternate path, thereby bypassing • Cyber security and data traditional legacy banks altogether. A vital characteris- • Pricing transparency tic of these collaborations is a sharing of each partner’s • Balancing integrity, innovation and a competitive SME digital data. This facilitates the development of new marketplace and innovative SME credit decision models and expanded access to credit. While a proportionate and enabling policy and regulatory framework will be need to support the use of alternative data • Access to data is no longer the problem in SME lending: for credit decision-making, it will also be important to im- Digital SME lenders have dispelled the long-held notion plement the various checks and balances. Addressing these that SME lending is not achievable in a scalable, efficient, issues will require: and profitable manner. In an increasingly digital econ- • A review of existing rules and regulations in place in order omy, these lenders are beginning to demonstrate that to improve the availability of reliable data for the purpose access to data is unlocking many of the earlier challenges to of enhancing SME financing, including access to bank expanding SME lending. The digital economy has also data transactional information. given rise to an ever-evolving set of value-added cloud- • Facilitating enhancements to improve credit information based services to help SMEs with their finances, business infrastructure, including SME credit reporting and access planning, productivity, legal issues, data backup and se- to data by alternative lenders. curity, file sharing, web conferencing, website builds, on- • Increasing cooperation of various regulators, not only on line marketing, business training, e-commerce, payments, a national level but also on a regional and internation- loyalty programs, business intelligence, and more. To al basis — especially to support innovative financing for increase customer engagement and help their SME cu- SMEs involved in global value chains as well as oversight of tomers be more successful, banks and other SME lenders alternative lenders operating in several markets. have started partnering with these platforms to offer SMEs • Involving policymakers in market competition rules. these applications individually, together, or wrapped up • Understanding the challenges and balancing act required with other core products and services. to address consumer protection, data privacy and the im- plications for increased cyber security measures in light of • However, access to data for SME lending brings new the use of new alternative data, new players and increas- challenges: With the abundance of alternative data, there ingly interconnected partnerships. are new issues of what to use, how to use it, and how to do this responsibly — while also respecting privacy and other important rights of SMEs. G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 1 Large SME Market, Credit Gaps, and Opportunities SMEs provide an extraordinarily large number of customers Deposit account balances and transaction flows provide re- and prospects for financial services in developing markets. al-time visibility into SMEs’ rolling net cash flows for credit SMEs number an estimated 400 million in the global de- decisions, which are often superior to out-of-date and un- veloping markets (Figure 1). Most (93 percent) are formal or reliable SME financial statements. They provide indicators informal micro firms. In addition, informal firms outnumber about incoming sales, outgoing expenses, debt payments, formal firms by a ratio of 3.4 to one. whether the business is growing or contracting, and how well the SME is handling overdrafts. About half of these SMEs (200 million) have unmet credit needs totaling approximately US$2.1 to US$2.6 trillion (Fig- SMEs in developing markets also spend an estimated US$8.0 ures 2 and 3). They also maintain an estimated US$14.6 to to US$10.0 trillion annually on business-to-business (B2B) US$17.8 trillion in cash balances (deposits), of which US$3.3 transactions (Figure 5). Digitization of these payments pro- to US$4.1 trillion are held outside of banks. They hold the vides valuable insights into SME supply chains for credit remainder, or US$11.2 to US$13.8 trillion, in current, savings, decisions. and investment deposit accounts (Figure 4). The size of the SME banking market is enormous. The con- sulting firm, McKinsey & Company, estimated that the global Figure 1: Global developing markets’ formal and developing markets’ SME banking revenue alone reached informal SME market size US$367 billion by 2015 (Figure 6).9 400 total 310 informal 40 The income mix is typically 40 percent deposits and 60 28 percent loans (including business credit cards and the as- 63 formal micros 1 9 14 21 sumption that overdrafts are included in the loan versus the 27 formal SMEs 2 4 deposit book).10 153 12 5 2 23 188 Leading SME banks can generate returns on assets (ROAs) 2 7 69 78 that are three times that of overall bank ROAs (three to six 3 percent versus one to three percent);11 SME banking returns 8 3 12 on equity (ROEs) of 25-33 percent;12 and SME lending profits 9 37 that are on average 35 percent higher than returns on overall 20 52 bank loan portfolios.13 Millions (estimated) 3 Sub-Saharan Africa Middle East and North Africa However, to achieve these returns, a sound, profitable SME South Asia Latin America and the Caribbean banking model must tackle the challenges of three key risks Europe and Central Asia East Asia and Pacific simultaneously – credit risk, excessive cost to serve, and 1. Registered firms with 1-4 employees lower revenue per account relative to large corporate clients. 2. Registered firms typically with 5 to 249 employees Since the majority of credit and deposit accounts are small, 3. Margin of error ranges from plus or minus 9.1% to 10.0% for each lending, relationship management, and distribution costs developing region, and plus or minus 9.4% for the global developing market total must be radically lower than corporate banking costs, and Source: IFC Enterprise Finance Gap Database (2011); Global Payments cross-sell of non-credit products are essential for driving up Experts LLC (GPE) analysis revenue per client, offsetting higher risk, and improving credit quality. 2 ALTERNATIVE DATA TRANSFORMING SME FINANCE Figure 2: Global developing markets’ formal and informal SME credit demand 400 = 100% 400 #SMEs, millions 1 350 188 300 250 200 20 182 = 45% 160 = 40% 160 = 40% 52 Unmet 150 88 credit 62 demand 77 100 78 12 8 6 25 22 18 21 39 = 10% 50 34 37 31 19 =5% 15 40 9 10 4 8 82 44 1 95 18 17 1 23 19 0 TOTAL Has checking Does not Has credit, Has credit, Unserved account need credit fully served needs more needs credit East Asia and Pacific Europe and Central Asia Latin America and the Caribbean South Asia Middle East and North Africa Sub-Saharan Africa 1. Margin of error ranges from plus or minus 9.1% to 10.0% for each region, and plus or minus 9.4% for the global market total Source: IFC Enterprise Finance Gap Database (2011); GPE analysis. Figure 3: Global developing markets’ SME credit demand Leading edge SME banks also carve out SMEs as a separate opportunity value stand-alone business with its own profit center, credit policy and lending ecosystem, unique target customer, value prop- $9,630 total 146 855 osition, and coherent systems of essential product, credit, $7,248 credit outstanding 15 496 575 and operational capabilities. They develop clear target mar- 368 $2,382 credit gap 132 ket sub-segmentation and simple, tailored product value 359 671 propositions, placing a high priority on the micro-small mass 706 1,291 market opportunity, and manage risk through an innovative, 4,774 207 strongly predictive, technology-led Retail credit-scoring 4,067 lending approach and risk-based pricing. Last, but not least, 358 620 1,631 successful alignment of SME relationship, credit risk, cost control, and credit policy dimensions requires sustained top executive support, commitment, and ongoing investment. 1,989 US$ billions (estimated) 1 Globally, on average, the SME banking market accounts for Sub-Saharan Africa Middle East and North Africa 27 percent of retail banking net revenues14 (this climbs to as South Asia Latin America and the Caribbean high as 50 percent when banks capture the owner’s house- Europe and Central Asia East Asia and Pacific hold banking relationship as well15). However, at most banks, 1. Margin of error ranges from plus or minus 9.4% to 10.3% for each much of this revenue today is concentrated in deposits, region, and plus or minus 9.8% for the global market total. Includes all formal and informal SMEs. cards, and merchant payment services; bank SME lending is often perceived as too costly and risky, making it difficult for Sources: IFC Enterprise Finance Gap Database (2011); McKinsey & Co. (August 2010), “Assessing and Mapping the Gap in Micro, Very Small, banks to pursue this opportunity. The SME segment is also Small, and Medium Enterprise (MSME) Finance;” “Two trillion and a large and heterogeneous group which is hard to assess. In counting;” GPE analysis. G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 3 Figure 4: Global developing markets’ SME cash balances Figure 5: Global developing markets’ SME B2B opportunity value spending opportunity value $16,207 total 211 106 967 283-343 $8,003-10,023 total $12,512 banked 177 75 841 541-649 629 34 32 689 $ 3,695 unbanked 1,419 338 807-970 151 1,192 227 3,236-4,050 2,014-2,587 2,912 9,751 1,122-1,425 12,663 US$ billions (estimated) 1 US$ billions (estimated) Sub-Saharan Africa Middle East and North Africa Sub-Saharan Africa Middle East and North Africa South Asia Latin America and the Caribbean South Asia Latin America and the Caribbean Europe and Central Asia East Asia and Pacific Europe and Central Asia East Asia and Pacific 1. Margin of error ranges from plus or minus 6.7% to 10.5% for each Source: Visa Commercial Expenditure (CCE) Index; GPE analysis. region, and plus or minus 10.1% for the global market total Sources: IFC Enterprise Finance Gap Database (2011); McKinsey & Co. (August 2010), “Assessing and Mapping the Gap in Micro, Very Small, Small, and Medium Enterprise (SME) Finance;” GPE analysis. addition, new provisioning requirements and Basel III add a executives worry that these challenges will lead to adverse further deterrent to regulated banks attempting to weigh the selection; top executives worry they will not earn sufficient potential benefits and costs of increasing SME finance. returns above their 10 to 15 percent cost of capital (Figure 7). Often banks that do provide SME credit insist on audited fi- In the wake of the global financial crisis, increased capital nancial statements, tax returns, and collateral requirements, and liquidity requirements, new regulations, and shrinking which most micro and small firms cannot provide. In ad- returns on equity, banks’ SME lending challenges are even dition, customer acquisition frequently involves using ex- more daunting. As noted, limited SME coverage by credit re- pensive brokers or middlemen,underwriting that is heavily porting service providers, weak contract or bankruptcy laws, collateral-based, and a tedious manual process requiring a and high SME informality in emerging markets compound lot of paperwork that often stretches to six weeks or more in banks’ challenges to offer SME finance. duration. Fear of rejection can also deter even credit-worthy SMEs from applying.16 All in all, banks in several markets face challenges address- ing SME financing needs. SMEs, for their part, are frustrated This in turn results in high costs to lend to SMEs. Given the with the entire customer experience and with banks’ lack of low revenue per account relative to large corporate clients, responsiveness to SME product and service needs. This cre- this limits most SME lending to the perceived safety and ates negative perceptions of banks which can be a broader comfortable margins involving the tiny segment of the big- issue than a lack of credit.17 gest of SMEs. This narrow view severely limits banks’ ability to capture the biggest slice of the SME opportunity. A World Economic Forum (WEF) report released in June 2015 predicts that incumbent financial institutions will come Bankers are trained to mitigate risk, not perpetuate it. Faced under attack in areas where the greatest sources of custom- with the challenge of managing a high volume of small cred- er friction meet the largest profit pools.18 Banks’ failures to its (typically, 80 percent of the loans generate 20 percent effectively serve SMEs have made them highly vulnerable to of the loan value) and credit information opacity, credit risk competition from the new breed of digital SME lenders. 4 ALTERNATIVE DATA TRANSFORMING SME FINANCE Figure 6. Three critical inputs – credit bureau coverage, SME client reach through branches, and the size of the SME banking revenue pool – help inform SME banking and lending business models and strategies % of Adults % share/amount of $367 2015 SME banking w/Credit Bureau revenue (billions) 50% Latin America/ the Caribbean 40% 16% Europe and Increased need for innovative risk models Central Asia $60 30% 13% East Asia $49 and Pacific 20% South Asia Middle East and 46% North Africa 17% 10% 4% Sub-Saharan 3% $15 Africa $67 $12 $167 0% 0 2 4 6 8 10 12 14 16 18 20 22 24 $ billions Branches/100,000 Adults Increasing need for direct channels to lower SME sales, credit, and servicing costs Sources: Mckinsey & Co. (August 2010), “Assessing and Mapping the Gap in Micro, very Small, and Medium Enterprise (MSME) Finance;” World Development Indicators. World Bank (as of 2014); GPE analysis Figure 7: Credit executives worry about adverse selection; top executives worry they won’t earn sufficient returns above their 10-15% cost of capital DEVELOPING MARKET SME CREDIT GAP $2.4 trillion1 200 million MSMEs1 SME DEMAND-SIDE ISSUES BANK SUPPLY-SIDE ISSUES • Difficult requirements • High risk • Cumbersome, slow applications • High cost • High costs/interest INFORMATION • Low revenue/account OPACITY • Fear of decline • Difficult to reach, dispersed • Low business, financial literacy • Informality Sources: IFC Enterprise Finance Gap Database (2011); McKinsey & Co. (August 2010), “Assessing and Mapping the Gap in Micro, Very Small, Small, and Medium Enterprise (MSME) Finance;” “Two trillion and counting;” GPE analysis. Margin of error is +/- 9.8% for the credit gap; 9.4% for # MSMEs. Includes all formal and informal SMEs. G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 5 The Rise of Digital SME Lending Fueled by digitization, the aftermath of the global financial All SMEs stand to benefit from access to digital SME finance. crisis, and the size of the SME banking prize, nimble, disrup- Medium and small SMEs will enjoy faster and better SME tive new entrants are unbundling SME banking relationships, lending decisions from a more comprehensive view of their nowhere more so than along the credit and payments value business, and many micro and informal SMEs will gain ac- chain. cess to formal credit for the first time through the enhanced visibility into their cash flows, payments, and trading as Digital SME lenders have begun to reach an important level well as innovations that grant credit based on the financial in several markets, especially given rapid growth rates. They strength and credit of large corporations that buy their prod- are no longer truly outside of the mainstream financial sec- ucts or services. tor; rather, they are beginning to represent a whole new way to offer SME finance. China SME lenders dominate the rest of While the rapidly unfolding digital SME finance landscape the world in terms of volume, reaching 8.6 times those of the constitutes a global laboratory, unique local market condi- United States (U.S.) and 23.4 times those of the United King- tions and digital data availability catalyze localized innova- dom (U.K.) in 2015. In the U.S., 2015 alternative SME lending tion solutions. This in turn flows back up to enrich the global volumes reached an estimated 4.3 percent share,19 whereas innovation laboratory. The future of digital SME finance no in the U.K., 4.3 percent of all new SME loans issued in 2015 longer flows just from West to East, but also increasingly were provided by digital SME lenders (Figure 8).20 from East and South to West and North; likewise, the flow between developed and developing markets is a two-way SME digital lenders are establishing themselves as serious exchange. Perhaps most interesting, developing countries’ SME finance players. While traditional banking players still leapfrogs to mobile have made their SMEs even more open dominate SME lending, the new digital SME finance lend- to technology adoption than their peers in developed econ- ers are growing rapidly. Unencumbered by legacy systems omies. Such rapid growth in SME digital footprints is likely to and processes, they are open to exploring cloud-based op- propel an acceleration in digital SME finance. Another factor tions and diverse digital data streams. They have engineered that has led to rapid growth and innovation in some devel- their solutions to begin to better address the credit needs of oping markets such as China or Kenya has been a lack of or SME customers today, as well as to reduce their risks. This weaker regulatory/legislative regimes and implementation includes a greater digital end-to-end solution, with front- capacity. However, there are also risks associated with lack end, simple online and mobile customer interfaces feeding of regulation with respect to the goals of financial stability digital back-end systems and processes with automated risk and a level playing field. analytics, as well as improvements in decision-making, client management and collection practices.21 Digital SME finance firms are also increasingly collaborating with each other, tech giants, banks, digital payment firms, SME-focused digital lenders are also beginning to drive and SME cloud service firms to expand into new SME lend- changes in SME customer expectations. Their simple and ing products (for example, adding invoice financing, lines of quick application processes allow SMEs to apply on any credit, or commercial property financing or adding personal device at any time, with much or all of the data collection loans or mortgages to serve the owner’s household), enrich automated. When faced with long-form bank applications their SME data streams through direct data sharing, and ef- that require the submission of paper-intensive documenta- ficiently reach new SME customers. In these ways, they are tion and an often multi-week decision timeframes versus the spreading a new generation of digital financing, business in- streamlined processes of the innovators, many SME owners telligence, and business development tools to SMEs across will choose the latter. the globe. In addition, even highly commoditized areas, such as credit scoring and payments, are facing growing 6 ALTERNATIVE DATA TRANSFORMING SME FINANCE Figure 8: Alternative SME lenders’ volume growth:1 top three countries $50 USD billions CAGR 2013- $45 2015 $41.66 $40 $35 $30 205% $25 $20 80% $15 60% $10 $8.44 $4.83 $5 $1.59 $1.78 $2.09 $1.47 $0.44 $0.83 $0 2013 2014 2015 2013 2014 2015 2013 2014 2015 U.K. U.S. China P2P SME Lending SME Balance Sheet Lending Invoice Financing 1. Excludes P2P SME real estate lending and digital SME lending by banks and tech, e-commerce, payments, and supply chain financing giants 2. CAGR = compounded annual growth rate Sources: “Harnessing Potential: The Asia-Pacific Alternative Finance Benchmarking Report” (March 2016), “Pushing Boundaries: The 2015 UK Alternative Finance Industry Report” (April 2016), Cambridge Center for Alternative Finance et.al.; GPE analysis Alternative SME lenders’ volume growth:1 other global regions $450 USD millions CAGR 2013- $400 2015 65% 376.8 360.4 $350 $300 $250 232.7 93% $200 186% $150 127.7 68% 116.6 83% $100 80.0 55.7 52.4 42.6 39.9 39.6 $50 6.9 15.1 11.8 9.9 $0 2013 2014 2015 2013 2014 2015 2013 2014 2015 2013 2014 2015 2013 2014 2015 Canada LAC Region Australia Asia Pacific Region Europe (Excl China/Aus) (Excl. U.K.) P2P SME Lending SME Balance Sheet Lending Invoice Financing 1. Excludes P2P SME real estate lending and digital SME lending by banks and tech, e-commerce, payments, and supply chain financing giants 2. CAGR = compounded annual growth rate Sources: “Harnessing Potential: The Asia-Pacific Alternative Finance Benchmarking Report” (March 2016), “Breaking New Ground: The Americas Alternative Finance Benchmarking Report” (April 2016), and “Sustaining Momentum: The 2nd European Alternative Industry Report” (September 2016), Cambridge Center for Alternative Finance et.al.; GPE analysis G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 7 digital competition. Further, new players continue to emerge predictive power of any SME scoring model, it is also im- globally. portant to note that not all the scoring technologies or data have the same predictive power across all situations. However, from a policy stand point, data fragmentation and While some variables commonly work across all markets, the potential for a few large players to “capture” or “control” there is no perfect “one size fits all” complete solution. The access to SME data could create monopolies and an atten- source, degree of strength, and quality of the data, ability to dant potential for anti-competitive practices that might lock capture the data in systems, credit bureau coverage, local in SME clients to a single lender. Over the medium to long- payment or business practices, cultural norms, and bank or term, policymakers and regulators will need to ensure that lender business lines, mix, and strategies are all factors that client data and access to credit information sharing are open can make data highly predictive in one market, lender, or and accessible to a broader range of SME lenders. situation and not at all in another. Digital data availability has largely been the driving force In addition, making use of alternative data requires putting behind the growth in alternative lending over the last few together exceptional and experienced teams of data scien- years. Digital SME lenders, including some digital banks, use tists and underwriters and developing a scoring technology new types of data for determining creditworthiness to target approach (particularly when working with unstructured data a certain segment of SMEs that are underserved by banks. like text from social media sites), that becomes progressively There is also evidence that they can analyze all of this data more precise and accurate as it draw lessons from steadily and issue credit much faster — and at much lower under- expanding data and underwriting decisions. writing and servicing costs than traditional banks. This is es- pecially true for those market niches that fall outside of the Another factor to keep in mind is that the earliest digital SME risk appetite for the more traditional banks. In these instanc- lenders were launched in 2006-2007; most have launched es, the value of speed and convenience is enough for SMEs only within the last five years. While a select few launched to take advantage of alternative lending models even when in the midst of the 2008 financial crisis and weathered well costs are similar.22 Further, the more diverse the data and the during the slow economic recovery period, no digital SME faster the analysis of the data, the more predictive its value lenders have been stress tested through a full economic (Figure 9). down cycle. In addition, just like online and telephone bank- ing, digital channels always attract a higher degree of fraud While powerful analytics and scoring of many different attempts and less creditworthy customers that all of these sources and types of data are a fundamental pillar of the lenders must guard against. majority of these SME digital lenders and can increase the For all of these reasons, it is important to stress market forc- es are still testing these business models and their use of Figure 9: More data, more diverse data = better risk models alternative data. While they show considerable promise and early proven reliability (where performance data is available), 100% the testing, learning, and adaption process is ongoing. As you move up in risk, you move up in default The rapidly evolving new generation of online finance com- 80% panies and digitized data make strict categorization of these Cumulative % bad accounts new business models difficult. Indeed, the lines between them are blurring as they collaborate with each other and 60% traditional banks, extend into new lending products, and respond to changing regulatory regimes. 40% There are four general institutional models using alterna- The area between the bowed tive digital SME data streams for SME lending decisions.23 In and straight 45-degree line = 20% addition, new “digital bank”-alternative finance partnerships improved ability to accept are also becoming more of a key trend. more good borrowers and reject more bad 0% • SME marketplace lenders broadly describe non-bank borrowers digital lenders originating loans to SMEs through intermedi- 0% 20% 40% 60% 80% 100% ary platforms, which connect SME borrowers to investors, Cumulative % good accounts rectly to their own balance sheets, or from a combina­ di­ Source: Global Payments Experts IIc. tion of the two. These lenders digitally access and 8 ALTERNATIVE DATA TRANSFORMING SME FINANCE substitute alternative SME data for missing conventional • Digital Bank models are among a few traditional banks, credit data, or enhance conventional credit data with al- as well as a host of new banks, that are directly develop- ternative data. As such, they can expand the pool of cred- ing their own in-house alternative lending systems. In this it-eligible SMEs and greatly simplify, speed up, and low- regard, they are opening up their APIs to third-party ser- er the cost of SME credit applications, underwriting, and vice providers, or acquiring or partnering with alternative portfolio management. lenders. • Tech, e-commerce, and payment giants are leading Leveraging data to remove friction between the customer global and/or country companies that have their origins in and the institution and improve risk management are at the online marketplaces, search engine providers, payments, center of these business models and companies. This sec- e-commerce, social networking, or computer technolo- tion reviews the diversity of SME alternative data used along- gy. They are leveraging their diverse, massive alternative side conventional data. It also examines the many forms of data streams generated from their own platforms or via collaboration and partnerships that are expanding the uni- their many partnerships. In this way, they are now offering verse of accessible SME digital data through representative loans and other financial services to their millions of cap- SME digital lender profiles. While selected company profiles tive SME customers. are used to best illustrate the data and its application to each business model category, their inclusion is not an endorse- • Supply chain platforms support SME financing during fi- ment or promotion of any company, business model, or the nancial transactions — purchase orders, invoices, receiv- reliability of a particular alternative data set. ables, other claims, as well as related pre-shipment and post-shipment processes — between buyers and sellers trading and collaborating with each other along the sup- 4.1 SME MARKETPLACE LENDERS ply chain. Triggers from the physical supply chain underpin each financial transaction. Emerging cloud-based digital Marketplace lenders are defined as intermediary online plat- supply chain platforms (including the use of blockchain24) forms, which connect SME borrowers to investors, SME to gain visibility and insights into the different parts of the lenders that fund loans from their own balance sheets, and complex trade flows between buyers and sellers (for ex- hybrid lenders that fund loans from both their own bal- ample, invoices, accounts payable, procurement data, ance sheets and from investors.25 This section illustrates historical business cash flows, shipping history, bills of the many types of SME digital data and credit approaches lading, economic indicators, and taxes paid). This is done emerging across this diverse SME marketplace lending sec- by digitizing documents and transactions and applying tor. It draws from a large number of: peer-to-peer (P2P) SME big data science and analytics to make credit decisions. lenders; online balance sheet lenders; payment innovators; They also leverage the financial stability and strength of analytic providers; identity information providers; cloud bond-rated, large corporates who buy SME products or -based lending platforms supporting lenders;26 SME loan services. As such, they can offer SME financing faster and broker marketplaces;27 and SME cloud-based, value-added often at a lower cost. products and services firms. Innovative supply chain financing platforms vary widely This section also analyzes how the various groups of players (for example, invoice or receivables discounting, payables use digital SME data, as well as how alternative data is ana- finance, dynamic discounting, working capital auctions, lyzed alongside conventional credit data. Another key and factoring, inventory finance, or pre-shipment finance), as growing trend concerns the partnerships developing be- do the funding sources for financing (for example, banks, tween fintechs themselves, as well as with banks. Together, retail and capital market investors, corporate buyers, they are working to expand both the universe of digital data and lenders that compete for the financing). For all con- available to lenders’ credit decision models and the pool of cerned, however, digitization provides for more efficient target SMEs that lenders can reach and serve. These new SME lending models. It helps suppliers, accelerates ap- lenders and partnerships are also better enabling SME lend- provals, increases SME credit access, reduces the chance ing across borders and increasingly around the world. of supplier or procurement fraud, and often, but not always, lowers the cost of financing for SMEs. 4.1.1 Peer-to-peer (P2P) SME lending platforms • Mobile-data based lending models offer small mobile P2P SME lending platforms originate loans for SME borrow- loans based on credit scores derived from mobile calling ers online for sale to retail or institutional investors. Investors patterns, mobile transactions, mobile e-money usage, can range from retail consumers and high-net-worth indi- and mobile e-money linked savings history, as well as viduals to institutional investors (for example, banks, hedge prior credit history data. G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 9 funds, pension funds, family offices, governments, and sov- transactions in lieu of providing bank statements or direct- ereign wealth funds). They typically focus on longer duration ly linking with banks. In addition, players like Faircent also loans, for example, one-to-five years. In this context, they tap into credit bureau data and identity information through collect upfront origination fees from borrowers and ser- their linkages with companies such as TransUnion. Access vice fees from borrowers, investors, or both throughout the to the Aadhaar-based eKYC service from TransUnion allows loan’s duration. Investors earn the interest spread on loans Faircent to streamline the process of identifying borrowers and bear the risk of loan default. Rates are often lower or and reduce the amount of verifying documents borrowers comparable to bank rates. are required to submit. Most platforms use automated credit screening models User Internet and loan portal behavior can also reveal valua- (some quite sophisticated) to risk, screen, and credit-rate ble insights. Lithuania-based SAVY evaluates the SME owner’ borrowers, set interest rates (or have investors compete in Internet behavior, history, previous loan requests and oth- an auction to set the interest rate within a floor or ceiling set er statistical data, for example, user behavior on the SAVY by the platform), collect loan payments, monitor and service portal, of the borrower, which can reveal the intentions of the loan portfolios, and handle collections and recoveries on the borrower. For example, the company has found that if delinquencies and defaults. the owner does not read information thoroughly and does everything quickly by using copy and paste functions, it Peer-to-peer SME lenders have focused on a range of means he wants to get the money as quickly as possible, and both traditional and alternative data, but their ability to is less likely to repay it. In contrast, if he spends more time manage this data in new ways is what has differentiated gathering information and modeling different scenarios, he them from other lenders. This includes the ability to auto- more likely plans repayment of the credit.29 populate underwriting information from a variety of data sourc- es to facilitate clients, as well as the ability to fast track credit In addition to obtaining information from banks, financial decision-making. institutions, credit reporting service providers, and online accounting systems, other P2P lenders are also obtaining The UK’s largest P2P lender is Funding Circle. It auto- information along supply chains. For example, South Afri- populates underwriting information through APIs from can-based Rainfin provides SME loans in a partnership with many data sources into its credit decision engine, allowing M2North, a company that enables SMEs and large industri- it to make very fast decisions. Although it looks at traditional al companies to exchange procurement documents. It also data —including business cash flow, personal cash flow, col- acts as an electronic intermediary between large companies lateral, and personal assets that could be liquidated if neces- and their supplier base. SMEs registered with M2North opt in sary - its focus on alternative metrics, such as real-time cash with Rainfin to share their existing data. This enables Rain- flow, Yelp reviews, and an owner’s active online engagement Fin to assess the credit-worthiness of a business, much like with the market, is also vitally important. performing credit checks — and provides a risk rating for the individual borrowers using its site. Like other P2P providers, it also obtains digital SME data, insights, and customers through its SME target and data-rich RainFin uses the data to calculate things such as a business’s referral partners, including Santander and the Royal Bank of estimated cash-flows. Also, since many SMEs using M2North Scotland (RBS) in the UK (banking and payment transaction have supplied large corporates, it can also access a firm’s data), software firms Intuit and Sage (cloud accounting and black economic empowerment status30 and value-added business financial data) and H&R Block (bookkeeping, pay- tax (VAT) registration. RainFin’s online application process roll, taxes and other accounting data). Partnerships with on- includes an intelligent SME-specific credit scorecard. This line SME loan broker marketplaces, such as LendingTree and scorecard reviews not only an applicant’s transactional his- Nav, as well as government and small business associations, tory and financial health, but also additional non-traditional help Funding Circle and other P2P lenders source potential data points. These can include procurement history and so- clients. cial media, and are used to assign the SME to one of seven risk-rating levels. Even where players are unable to directly access data from linkages to banks, alternative approaches include the use Other examples of innovative partnerships and analysis of screen scraping28 to gather data from banking and pay- by P2P lenders and supply chains comes from US-based ment transaction data. India-based Faircent integration with Lending Club. It is partly fueling its SME loan growth through Yodlee provides it with bank scraping technology. This en- major SME lending partnership deals struck with Google ables Faircent, with SME permission, to digitally link and and Walmart’s Sam’s Clubs. This gives Lending Club ready update multiple borrower deposit and credit bank account access to millions of SMEs and their valuable data virtually 10 ALTERNATIVE DATA TRANSFORMING SME FINANCE for free. For example, Google uses Lending Club to provide and a good payments history on PayPal is unlikely to act low-cost, two-year loans of up to US$600,000 to its Google fraudulently. for Work network comprised of more than 10,000 partners. Google already knows the borrowers and provides data (for P2P SME lenders focusing on supply chains include compa- example, data on sales, contracts, income, and identification nies such as India-based LoanZen. It caters only to private/ information) to Lending Club. Lending Club in turn analyz- limited company borrowers in services and manufacturing es this data to evaluate Google SMEs’ creditworthiness and that have large, reputed clients (that is, multinational cor- services the loans. porations [MNCs] or listed Indian corporates). SMEs across these sectors furnish their invoices on the platform and re- US-based ApplePie Capital, which lends solely to fran- ceive unsecured loans from accredited investors who are chisees looking to start, grow, or retrofit their businesses, comfortable with the borrower credit profile. Loanzen then obtains data from its franchisor partners (such as branded enters into a tripartite escrow agreement administered by gyms, beauty parlors, pizza delivery, or quick service restau- a bank to collateralize the receivables. The invoice is not rants). Franchisees follow proven business models backed “sold;” rather, LoanZen uses it primarily to determine the by the central resources of a national or regional franchisor volume and tenor of the expected cash flows against which brand with strong incentives to see their franchisees suc- to extend unsecured credit. Borrowers connect their ac- ceed. Franchisor data includes such things as how many counting, tax, and online banking data to LoanZen’s artificial stores they have opened or closed over what period of time, intelligence-based system, which completes the credit as- what a store should make in volume or revenue, what it costs sessment within 15 minutes. to build, how long it takes to break even, and other unit eco- nomics broken down by geography and store footprint size. LoanZen has also partnered with other companies to sup- The franchisor also provides the criteria it uses to select fran- port loans to numerous linked companies such as Treebo, chisees (ApplePie’s loan customers) and the market demand a technology-and-analytics-enabled hotel chain. LoanZen size for the product or service in the franchisee’s store lo- uses Treebo’s data on their partner hotels. Such data in- cation. Feeding this kind of predictive alternative data into cludes past booking history, future bookings for the prop- its credit underwriting and credit line size models enables erty, guest feedback collected digitally, and quality perfor- ApplePie to better underwrite loans to franchisees. ApplePie mance data. As such, they can then offer credit to a segment also leverages its partnerships with brand name franchisors of small hotel owners that has so far not had access to such for low cost borrower acquisition.31 lending. Other innovative partnerships that result in new alternative Individual behavior, as well as the social media connections data come from e-commerce platforms. Alibaba’s e-com- of small business owners, is seen as relevant. Faircent (with merce platform, Alibaba.com e-Credit Line, is powered by borrower permission) collects alternative data from social Lending Club and demonstrates the importance of collab- platforms, such as LinkedIn and Facebook, to supplement oration. Lending Club gets direct access to Alibaba’s.com its analysis. The platform’s algorithms also detect good and large base of U.S. buyers and sellers as well as their online bad credit behavior. For example, a borrower struggling to transaction and trade data with Chinese suppliers to more pay back his debt after spending recklessly and hitting his accurately assess risk, and make faster decisions. SME loans maximum credit limit will rank poorly on the site. Faircent can made entailing less risk, and at lower interest rates. With also evaluates borrowers’ lifestyle patterns and how he or direct access to Alibaba’s SME trade and sales data, Lending she spends money (for example, by buying the latest phone, Club can match repayment terms to the cash-flow cycles or frequenting a pub, among other things) by analyzing the of borrowers, vet suppliers and shipments, and transfer the bank account and payment transaction data collected. funds directly to the suppliers. Lending Club is fully respon- sible for underwriting. The risk is then borne by a pool of In addition to social media, P2P SME lenders are also ex- institutional investors, which gain a new low-risk asset class ploring the use of psychometric credit information for credit for investment. scoring and screening.32 Indonesian-based Moldaku ana- lyzes potential borrowers by going through five steps in Another example of e-commerce analysis by P2P SME lend- the screening process, which includes a profile screening, ers includes German-based Bitbond. By allowing borrowers an anti-fraud verification with a site visit, and a psycho- to connect their eBay and PayPal accounts (among others), metric credit information tool test facilitated by financial Bitbond gains an understanding of the creditworthiness of technology company Entrepreneurial Finance Lab (EFL). the applicant. Specifically, an online seller with a careful- With EFL, applicants fill out a 25-minute questionnaire ad- ly guarded reputation, large amounts of positive feedback, ministered by the lender, and in less than 10 minutes EFL G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 11 generates a credit score. The score is based on the appli- online, real-time credit for sellers on partner e-commerce cant’s answers to questions capturing information that can marketplaces, including Amazon.cn, Alibaba’s Lazada, Tao- predict loan repayment behavior, including the applicants’ bao, Tmall, and AliExpress platforms, eBay, and Wish. Sellers attitudes, beliefs, integrity, and performance. EFL then authorize access to their ecommerce data, including trans- analyzes the data to produce a credit score that assesses action volume, shop size, customer comments and ratings, the applicant’s ability and willingness to repay a loan in real and business performance, as well as other third party data time. EFL also uses alternative data such as psychometrics, such as purchase orders saved in enterprise resource plan- digital footprints and cellphone usage information to assess ning, logistics, and inventory management software or ser- the repayment risk profile associated with any individual. EFL vice providers they use. ShangTonDai also uses public data continues to improve its psychometric credit scoring capa- about the seller on the internet and social networking sites bilities, while simultaneously innovating with new alternative to make the lending decision. data sources. These sources include mobile phone usage data (through call detail records [CDRs]), social network data ShangTonDai is part of the YiQiFin platform for SMEs, po- (through Facebook and Twitter, for instance), and location sitioned as a “Cloud CFO” focused on enhancing clients’ data (through the Global Positioning System [GPS] and the business capacity across a range of functions for payments, Geographic Information System [GIS]). wealth management, internal financial management, and financing processes. The platform integrates elements of 4.1.2 China’s P2P lending platforms specific industry structures, supply chain, the internet and finance, and includes SME services such as credit analysis China’s P2P lending platforms are different from their and working capital gap calculations from upstream and counterparts in other markets due to their unique downstream suppliers, incorporating inter-bank and the offline-to-online (O2O) feature. In China, many small firms third-party payment options so that clients can operate cash are informal and accounting records are not often availa- flow and manage budgets, and wealth management. In the ble, borrowers are harder to vet. As a result, many Chinese agriculture segment, CreditEase lets SMEs and farmers lease platforms employ a hybrid O2O model: platforms source rather than purchase small equipment and over 180 kinds of investors online, but do customer acquisition and credit agricultural equipment, and even includes a livestock leasing and background checks on borrowers offline. They achieve program that lets ranchers rent cows. Most recently, in May this by partnering with non-bank financial institutions or by 2017, CreditEase launched a joint venture with Tradeshift, the platform’s own agents or staff. However, this also adds one of the world’s largest business commerce platforms, to considerable expense. These include visits to the borrower’s deliver a trade financing app that will bring low-cost financ- place of business to take pictures of the workplace. Some ing to millions of businesses in China. The joint venture will platforms in other emerging markets with similar profiles be extended from the electronic invoice to the upstream and have also incorporated some offline borrower authentica- downstream supply chain processes, creating a closer trade tion steps, primarily to meet Know Your Customer (KYC) re- link for global sourcing and supplier interconnection. quirements — such as checks on physical identification (ID) cards. Consumers in China have generated a tremendous amount of data, even if they lack a proper credit history. These in- One of the earliest P2P lenders, CreditEase has become a clude data from online search, social networks, online shop- leading fintech company in China specializing in small busi- ping and payments. With data from the digital footprints of ness and consumer lending and wealth management for high potential clients, P2P lenders like China Rapid Finance (CRF) net worth and mass affluent investors. Its cloud-computing use proprietary technology to establish credit scores for Big Data Innovation Center (BDIC) infrastructure, among them. other things, powers all credit operations for the company, enabling the company to tap e-commerce, telecommuni- Like CRF, China-based P2P lender Dianrong has also un- cations, bank and credit card, insurance, supply chain, and derstood the importance of tapping into and leveraging social security data for its SME and consumer credit scoring e-commerce and online activities of potential borrowers. and lending decisions. CreditEase also relies on these data For example, Dianrong uses data from Ant Financial’s cred- sets to grade borrowers for risk, and then apply risk-based it scoring service, Sesame Credit. It also links and analyzes pricing based on the borrower’s risk grade. information from potential clients’ social media accounts, such as Weibo, China’s equivalent of Twitter. Apart from ana- Using its ten plus years of accumulated credit, fraud, and lyzing social media usage, clients who miss a payment may other data assets, the company has also created customized potentially be embarrassed online because Dianrong.com products for specific SME segments. ShangTongDai powers can post public requests on social media site to demand 12 ALTERNATIVE DATA TRANSFORMING SME FINANCE recovery.33 In addition to Alibaba, Dianrong is also partner- Merchants can optionally link their Facebook, Twitter, and ing with eBay to lend money to Chinese businesses that sell UPS shipping accounts, which may qualify them for fee goods to U.S. customers on eBay. More recently, Dianrong discounts. Executives describe this data as the ‘space between partnered with FnConn34 to launch Chained Finance, a new that data’ to decide if Kabbage is going to offer the merchant blockchain platform for supply chain finance. The platform capital or not. Facebook business, Yelp, Foursquare, Amazon, records and authenticates every payment and every supply and eBay offer business reviews, rankings, and other rich data chain transaction, creating greater visibility into suppliers and regarding how SMEs actually interact with their customers. their trading data for SME lending decisions. For example, Kabbage has determined that customers who link to their social media information are 20 percent less like- 4.1.3 Online SME balance sheet lenders ly to be delinquent than those who do not. The company has also found customer review data – how long the company SME balance sheet lenders retain the portfolio and collect has been receiving reviews, what the trajectory has been for the interest rate spread over the lifetime of the loans. This volume and quality over time – is predictive.36 increased return and steady cash flow comes with the risk of possible loan defaults. Many balance sheet lenders tend to Similarly, linkages with logistic and e-commerce providers focus on specialized lending, such as riskier newer or small- are also producing relevant data. UPS shipping data can er SMEs, point-of-sale (POS) financing loans, merchant cash reveal how many packages an SME is shipping, how many advances, or factoring; others handle the full credit spec- packages are returned, the longevity of the business, and if trum. In general, the duration of these loans is shorter (for the quantity of packages shipped is going up or down. If the example, ranging from one month to three years) and they company knows that someone has been shipping antique have higher rates than those of P2P SME lenders. mugs for at least two years for eBay and Amazon, always ships out via two-day UPS air, has more than 500 friends on They generally mine transactional data as a key source of Facebook, and is always sending out deals on Twitter, then information in their credit scoring model. In addition to pro- they are often a better risk regardless of the credit score.37 viding loan origination and risk-based loan pricing, they are also able to reduce the costs and risks associated with repay- Online SME balance sheet lenders also take advantage of ments by automating loan payment deductions from busi- analyzing e-commerce transactions and linking with play- ness accounts or directly from regular sales transactions on ers such as Malaysian-based GAX Finance. It partners with a regular basis, most often daily. Transactional data also sup- e-commerce portals such as Lelong.com.my and auto re- ports ongoing risk management, and provides continuous pair workshops to tap into their data. It is also extending its feedback on the situation of a borrower using data analytics. solutions to medium-to-large business-to-consumer (B2C) Most also have a strong bank channel, wholesale funding, retailers with payment gateways, as well as SMEs offer- servicing, merchant acquiring, and/or data sharing and cus- ing software-as-a-service. UK-based iwoca allows SMEs to tomer acquisition distribution partnerships. These lenders upload and link their e-commerce accounts (for example, have also been the most active in bank-digital SME lender eBay, Amazon, notonthehighstreet.com, and ekmPower- collaborations. shop). E-commerce accounts show sales, active product listings, and feedback ratings. US-based Kabbage typically analyzes large numbers of transactions for each loan. It achieves this by pulling al- India-based Capital Float’s partnerships include B2C e- ternative data from its customers’ bank accounts and/or commerce platforms Snapdeal, Shopclues, Paytm, Flipkart, merchant accounts, merchant acquiring processors, social and Amazon as well as B2B e-commerce sites Alibaba.com, networks, e-commerce sites, accounting software, ship- Tolexo, IndustryBuying, and OfBusiness, among others. The ping records (United Parcel Service [UPS], for example), and company offers loans to SMEs on these platforms, and SME dozens of other private and public sources to gauge the risk loan applicants can digitally upload their seller e-commerce and creditworthiness of the business seeking the loan. It also trading to Capital Float. Capital Float’s scoring technology pulls updated information from these data sources daily. also evaluates merchant customer reviews on social media A typical Kabbage loan transaction, say executives, could call and e-commerce sites. In addition, it does psychometric on 30 different data sources and up to 50 different models to assessments (for example, applicants are asked questions score that data at any different time.35 The company’s under- about their ability to scale a business, attitude toward credit, writing engine pulls information such as business revenues, and how they compare to competitors.) vendor payments, and tax and accounting data to assign and adjust the proper line of credit in real time. Capital Float overlaps alternative digital data with government data like Aadhaar a identity data and the SME’s profitability G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 13 parameters, such as current product usage levels, industry (GooglePlay and Apple) and analytics accounts (for example, margins, potential future earnings, and risk parameters (such Flurry, Localytics, MixPanel, and AppsFigures), which allows as probability of default). It combines this with conventional it to quickly analyze past performance and future potential. credit bureau metrics, and tailors its credit-scoring model for Using their algorithms, Aprenita can evaluate a company’s each category of potential borrower it serves. For example, creditworthiness within a few minutes and issue funds within many of its customers want to grow businesses they have 24 hours. Data Aprenita accesses in the mobile apps market built on e-commerce platforms, whereas others are plan- includes customer engagement, sales, marketing conver- ning to sell online for the first time and need startup capital. sion, outstanding invoices from the app store, and customer In this regard, small-scale manufacturers have different pro- feedback, among other data. The app stores also provide files than retailers. information about the number of times the app has been downloaded, as well as reviews by customers. Advertising Capital Float has recently launched a mobile loan application networks provide data on the amount of revenue the app is which can approve loans in less than eight minutes. Within a generating. Other data analytics platforms reveal how many few months, the company is already seeing 50 percent of its active users the app has, how much time users typically applications and online browsing come in through its mobile spend on the app, and data on the effectiveness of the app’s platform. It has also struck a partnership with Payworld, a marketing efforts. payment innovator targeting customers in remote locations, to reach India’s 12 million kirana, or local neighborhood POS and other merchant-based sales data are also being stores. In addition, Capital Float has built a customized credit harnessed by marketplace lenders to facilitate access to model to provide credit to these retailers. credit. US-based Square Capital utilizes machine learning models. It identifies and makes offers to growing businesses Another example of alternative data comes from receiv- it deems credit-eligible based on the SME’s sales and pay- able-based invoice financing and analysis of SME’s online ments growth data, the mix of the SME’s new and return- account data. Australia-based Waddle provides automated, ing customers (an indicator of how the company grows), receivables-based invoice financing and a financing add-on the daily number of and size of sales tickets, and cash flow, that uses the SME’s cloud accounting (as well as bank ac- among other information sources. The loan functions sim- count information) to make SME loans unusually simple. It ilarly to a merchant cash advance, which is a sale of future is a fully online, cloud-based platform enabling SME owners receivables. However, it is technically structured as a loan to obtain automatic approvals (automated real-time lend- subject to lending regulations that must be paid off within 18 ing) and ongoing revolving credit lines based on outstanding months of acceptance. invoices held in their  online accounting packages. Funding occurs within 24 hours. The credit line is only for SMEs that Iwoca also uses online payment or POS merchant accept- transact B2B. ance accounts (for example, Magento, Skrill, Shopify, Sage Pay, Paypal, and Linnworks), as well as online accounting Waddle links directly into accounting and banking data, al- (for example, FreeAgent, Sage), business bank statements, lowing it to provide revolving credit lines to close cash flow VAT returns (which can be downloaded directly from the gaps to better support business growth. Once the account- U.K. government’s Her Majesty’s Revenue and Customs web ing application is linked, Waddle calculates a “borrowing site), and company accounts during the application process. base” (the total amount of eligible collateral) based on the VAT returns provide sales history; company accounts show business data. Waddle then establishes a fluid line of credit business profitability; payment and POS accounts capture to the business. The more the borrower uses Waddle, main- sales and identity information; and accounting records pro- tains an excellent repayment history and demonstrates high- vide a comprehensive view of the business financials. er sales transactions, the higher the credit limit. Waddle also integrates with Xero, MYOB and QuickBooks Online. By lev- Similarly, Capital Float uses SME mobile POS payment pro- eraging API technology, Waddle has the ability to automate viders MSwipe and Pine Labs, as well as information from the entire lending process. Business owners that use cloud transportation-sharing service Uber (vehicle loans for accounting are able to link their online accounts to Waddle, drivers), and cloud accounting software provider Intuit. and can opt-in for a two-way data exchange automating Kenya-based Kopo Kopo Grow Cash Advance grew out every aspect of financing. of an effort to support SME mobile money payments. This occurred after it partnered with Safaricom to acquire mer- With a focus on lending to the growing mobile app devel- chants to accept M-PESA at the point of sale (this service oper niche, US-based Aprenita sources alternative data on was later branded Lipa na M-PESA, Swahili for “Pay with their borrowers through direct integration with App Stores M-PESA”). 14 ALTERNATIVE DATA TRANSFORMING SME FINANCE The company launched Kopo Kopo Grow Cash Advance demonstrate that it can be introduced with a high degree of in Kenya, a merchant cash advance product for its mobile success.38 money merchants. The company’s merchant payment data enabled it to build a credit profile that analyzes over 200 In emerging markets struggling to increase payment ac- variables to price risk and to extend unsecured loans. The ceptance points, there is strong evidence to demonstrate product crunches hundreds of data-driven “signals” to pre- that a working capital product can be a powerful element dict a merchant’s future cash flow and propensity to default. not only in building a more compelling value proposition for It then pre-qualifies that merchant for loan ranges tailored merchants, but also singularly effective in growing merchant to the business. The merchant then selects the loan amount payment acceptance. Merchants are highly motivated to desired, dedicates a percentage of daily mobile money sales convert their customers from cash to electronic payments to repaying the loan, and digitally signs the terms and condi- in order build the electronic transaction history that will tions. The whole process from application to loan disburse- qualify them for credit access. For example, a Kopo Kopo ment can take minutes. Kopo Kopo automatically deducts a analysis that compared merchant sales during a period three percentage of every single mobile money payment in order months prior to accessing the loan (reflecting the incentive to amortize an outstanding loan. As a result, merchants do to achieve a good first credit score), and three months af- not have to remember to pay installments over the term of ter (reflecting the incentive to repay quickly and earn bet- their loan. It gets repaid automatically, transaction by trans- ter terms), showed a 42 percent higher transaction growth action, every single day. among Grow Cash Advance merchants.39 Zoona is a top mobile money operator in Zambia. The Another group using similar data are the hybrid funding SME company launched Zoona Growth to provide an affordable platform lenders that combine funding elements of both and accessible working capital financing package for Zoo- SME balance sheet lenders and P2P business models. These na agents. It is linked to customer usage and the growth of lenders fund loans from their own balance sheets as well as Zoona payments. Zoona agents may pre-qualify for the from investors. They lend their own capital directly to SMEs, product and can access larger facilities as their payment but also access institutional investor capital or securitize volume grows; it is particularly popular among rural agents. their loans to fund a portion of their portfolios. In terms of al- Due to its relationships with its agents and mobile transac- ternative data usage, both online SME balance sheet lenders tion data, Zoona has the ability to carefully perform credit and hybrid funding SME platform lenders gather similar data. scoring for individuals and manage default risks. Like online SME balance sheet lenders, hybrid SME lenders Australia-based Tyro Payments, an electronic funds transfer are using similar alternative data to support lending models. at point of sale (EFTPOS) provider, offers Tyro Smart Growth US-based OnDeck is the leading online SME balance sheet Funding as an SME financing service for its merchants. The lender with hybrid funding. According to OnDeck’s propri- financing is based on the company’s cash flow, its financial ety credit scoring model, it analyzes more than 2,000 data health (as seen through the data streaming in from Tyro’s points from over 100 external data sources and 10 million POS), and cloud accounting tools linked to Tyro and already SMEs in its proprietary database. It then creates an SME in use by the business. business credit profile. This data ranges from cash flow and transactional data to public records, as well as its own exten- UK-based Worldpay is another global payment processing sive internal historical performance data. provider. It now shares its data with alternative lenders to offer merchant cash advance services. It teamed with US- Similar to examples of other marketplace SME lenders, based CAN Capital to allow merchants to apply in minutes companies like OnDeck also facilitate the credit application using a simple and fast application process online or over process. It integrates with online accounting and banking the phone. Merchants can receive funds in as little as two transactions as well as POS merchant data to better analyze business days. Payments are based on a percentage of daily data, such as transaction frequency and volume, season- card revenues that are automated as daily remittances from al sales, expenses, and customer revenues. The company the borrower’s merchant account. then analyzes personal and business credit histories from credit reporting service providers. It also scans public and What these profiles clearly show is that SME cash advances legal records for past lawsuits or liens, and reviews Occu- based largely on the SME’s own merchant payment histo- pational Safety and Health Administration (OSHA)40 records ry are a product that is equally viable across both develop- for violations. Finally, it considers the health of an applicant’s ing and developed markets. In this regard, it can be slight- industry and region, and checks online business reviews ly adjusted for local market conditions. These cases also from sites such as Yelp, Angie’s List, and Google Places. In G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 15 order to take into account different types of SMEs, OnDeck mobile penetration, as well as accommodative regulations, applies one of a dozen different statistical models to the data have helped to fuel their growth.41 depending on business age, industry and geography. The Alibaba e-commerce giant’s domestic and international However, unlike the other marketplace SME lenders, virtual marketplaces and its financial services arm Ant Finan- OnDeck is also able to underwrite a wider range of credit cial serve over 425 million active buyers, over 10 million ac- products, which gives it a cost advantage in acquiring cus- tive SME sellers, and over 450 million active Alipay payment tomers. OnDeck can now: underwrite a short duration loan users.42 Ant’s MYBank, an internet-only bank, launched in to a relatively new business; provide a line of credit prod- June 2015 and in early 2016 merged with Ant Micro Loans, uct to a business with sporadic cash flow needs; or arrange Ant’s SME lending arm, which provides loans to Alibaba’s a bank-like, multi-year loan to a mature business. OnDeck merchant sellers up to $155,000. Ant Financial has devel- receives automated electronic repayments from its borrow- oped Zhima Credit (Sesame Credit), a credit rating service, ers either on a weekly or daily basis, giving it extensive in- which leverages “big data” technology and customer behav- sights into its borrowers’ cash flows. This feature enables it ior analytics. It uses both online and offline data to generate to quickly offer loan adjustments to those borrowers having credit scores for consumers and SMEs. In addition to owner trouble — or to revamp its model if a batch of loans begins characteristics, the company takes into consideration: the to experience difficulties. records on sellers, including the number and value of their sales; their cash flows through Alipay; comments posted by their buyers; tax payments and customs declarations for us- 4.2 TECHNOLOGY, E-COMMERCE, AND ers who export; shipping and logistics data; and even utility PAYMENT GIANTS bills from sellers’ factories. Indeed, vendors that have been in business for only two or three months can secure a credit The world’s digital giants now moving into financial services line using this wealth of real-time data. and SME lending have their origins in online marketplaces, search engines, payments, social networking, e-commerce, Alibaba and Ant Financial can also spot vendors who have or computer technology. What ties them together is that been too aggressive in certain fields and lagging in others. they already have millions of captive SME customers , began They achieve this by evaluating a mix of data, including their as all-digital companies, and use digital means and big data promotional campaigns and profit margins. Based on the to offer new financial services, with a notable focus on SME results, Ant can provide suggestions for how the vendors finance. need to adjust their operations, and provide financial sup- port accordingly. Another algorithm allows Ant to pace lend- Compared to other alternative SME lenders, they control ex- ing more effectively by increasing credit lines to accelerate tensive data about SMEs and the customers they serve and inventory purchases needed for big promotions later in the can allocate large investments in developing their financial year. This can well exceed the typical lending maximum of services initiatives. They strive to offer top-notch digital US$155,000 for qualified merchants.43 services and experiences. Indeed, they are far ahead of traditional banks in using analytics and artificial intelligence In January 2016, Alibaba forged more than 25 partner- to understand customer preferences and behaviors. ships with credit rating agencies and financial institutions in China and other parts of the world. These new partnerships 4.2.1 CHINA are enabling Alibaba to better offer SMEs cross-border trade finance, as well as to enhance their credit rating scoring tool China’s national champions in e-commerce (Alibaba and for SMEs. The service may also help overseas buyers identify Ant Financial), social media (Tencent), and search engines trustworthy trading partners and provide Chinese suppliers (Baidu) have all made concerted moves into finance with access to even more financing options.44 private banking platforms. They may later outdo anything we have seen in the West to become the world’s largest center Tencent is China’s leading social media and gaming platform of digital banking. These new entrants were faster than the in China. It maintains ubiquitous messaging platforms, in- banks to offer convenient, reliable, fast and cost-efficient al- cluding Weixin/WeChat messaging (752 million active users) ternatives to traditional bank payments. Their client numbers and QQ mobile messaging (877 million active users), as well match or exceed China’s top banks. Further, they have more as its Qzone social network (648 million active users). It also financial resources at their disposal, which means that they owns the number two Tenpay payment platform in China, can sustain larger, more balance sheet intensive businesses with over 300 million Weixin/QQ accounts linked to bank than their Western counterparts. High national Internet and cards for payments.45 In January 2015, Tencent launched 16 ALTERNATIVE DATA TRANSFORMING SME FINANCE China’s first online-only private bank, WeBank, in a joint supply the loans from their own balance sheets. The banks venture with two investments firms and several other share- bear the risk, and own the loan assets. holders. The eBaotong loan (a near instant loan offered in conjunc- Tencent also announced it would launch the Tencent cred- tion with China Construction Bank) is a typical offering. The it rating tool to analyze SMEs. It has partnered with China seller needs to arrange the product shipment first, and after Rapid Finance to help crunch the Tencent’s massive data entering the tracking number on DHgate.com, a dialogue troves to create credit ratings for some 800 million Tencent box will pop up immediately on the screen, asking whether users.46 Tencent also uses data from its many partner com- the seller wants to apply for a micro loan. After receiving panies to assess SME credit. For example, WeBank’s inaugu- customer authorization, DHgate.com supplies partner fi- ral loan to a Shenzhen trucker was based on data provided nancial institutions with the big data generated on DHgate. by a Tencent-invested logistics platform called Huochebang, com so that they can analyze it and determine risk. Sellers or Truck Club. Huochebang’s app links logistics providers can often receive a micro loan within 30 minutes for an with truck driver companies that need to ship cargo. As of amount up to 80 percent of the goods’ value. September 2014, its platforms were serving 167,000 logis- tics customers and nearly one million drivers with 650,000 The data generated is not just limited to a customer’s trans- trucks. Huochebang’s large data bank contains information action history; buyer feedback, logistics data, and inventory about each club-member trucker, such as total travel dis- data all factor into risk assessment. Data analyzed includes tances, what kinds of orders have been handled, as well as factors such as: the average number of orders per month; cargo volumes. WeBank was able to leverage this data to the total transaction amounts per month; and the number provide the loan to the Shenzhen trucker to pay in advance of disputes received. It also includes the duration of being an for the freight being hauled.47 active seller, number of consecutive transaction days, date of first order, buyers’ loyalty to this seller, return, dispute, and Baidu is the leading Chinese language Internet search pro- loyalty rates, and more. vider. In November 2015, Baidu established Baixin Bank, a direct bank, in a joint venture with China CITIC Bank. Baixin After a loan is released, the after-loan management data Bank will have the advantages of both Internet-based oper- analysis is used to detect abnormal behaviors, monitor the ations and the convenience of traditional banking accounts. process, predict trends, make related comparisons, and alert Baidu will bring its user traffic resources, behavioral data financial institutions if there are identified risks. DHGate’s on users and data analytics to Biaxin Bank, and CITIC will system also ranks specific performance measures, such as bring its knowledge of financial products. Although much of sellers’ response time to questions, product quality, product the initial focus was on consumer financing opportunities, information, and the information sellers provide on the ship- China Rapid Finance’s analysis of search engine data from ping status of orders, thereby making them more reliable Baidu reveals ample opportunities to connect to potential predictors of business risk and stability. small business borrowers. The data reveals correlations between search engine phrases and borrowing needs and 4.2.2 THE UNITED STATES behavior. For example, small business owners that search keywords “photography” or “hiking” are likely to need an American tech companies — Amazon, Apple, Facebook, and unsecured small loan. They are also likely to be looking for Google — entered the financial services market a few years their first loan, as compared with individuals searching key- ago. While all offer payments services, two of them, Am- words such as “cars” or “scuba diving.” However, small busi- azon and Paypal, also offer SME lending to sellers on their ness owners searching online for “lottery tickets” are unlikely platforms. to be a safe bet for lending platforms.48 Seattle-based Amazon, which has over 300 million active DHgate.com is the biggest transactional cross-border B2B users and 176 million site visits a day, began offering short- e-commerce marketplace in  China. DHGate began of- term inventory financing loans (of three-to-six months) to its fering SME financing in 2010 in partnership with leading some two million SME third party sellers beginning in Octo- Chinese banks.49 DHfinet, its internet finance branch, sup- ber 2012 in the US, the UK, and Japan. Since Amazon moves plies e-commerce micro-loans to SMEs so they can scale merchandise for e-commerce vendors, it has information up their businesses. SMEs applying for the loans must con- about what products customers are buying, what products duct business on DHGate’s cross-border e-commerce are shipped and returned, as well as statistics on payment platform, which generates a tremendous amount of data. trends. It also relies on close relationships with sellers and DHgate.com then partners with financial institutions to mitigates risk by taking loan payments from proceeds due to G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 17 sellers for their sales.50 A business must sell a certain volume lack of formal banking relationships, no or poor proper doc- of goods on the Amazon storefront before it can receive an uments and financial statements. At the same time, many offer of a loan. In order to qualify for such financing, Amazon SMEs in India have embraced the digital world by using mo- also focuses on data related to the merchant’s reputation bile phones and payments, social media, and e-commerce with its customers. (as both sellers and buyers). US-based PayPal Working Capital offers SME customers Digital SME lending in the Indian context, not unlike the flexible working capital advances for a risk-based flat fee, Chinese context, aims to use digital data as a surrogate for no additional costs or penalties, and no time limit on repay- financial and missing credit history data. With the SMEs’ per- ment. PayPal “members” — those who have been with the mission, the new SME lenders farm digital data from banking payments firm through a business or Premier account for at and payment transactions (including deposits, cards, current least 90 days — are eligible to apply. The company automat- accounts, mobile usage, people-to-business [P2B] busi- ically deducts repayments as a fixed percentage of daily sales ness-to-government [B2G], and business-to-business [B2B] (for example, the business pays more when sales are strong transactions), financial and business data in cloud account- and less when sales are slow). ing and customer relationship management (CRM) systems, government registration, identity, and tax records, and credit Business owners choose the percentage to deduct when reporting service providers. they apply. The amount ranges from 10 to 30 percent; the higher the percentage, the lower the flat fee. Paypal draws Fast-growing B2B and B2C e-commerce platforms can pro- on insights into how its merchant customers operate, and vide rich sources of buyer and seller trading transactional uses its sales history data to power rapid lending decisions. information, information on sales and returns, and custom- Poor eBay seller ratings, a history of chargebacks or too er reviews and ranking data. Large corporates’ data on their many chargebacks, active PayPal disputes, holds on the SME suppliers (procurement, purchase, invoicing, order ful- PayPal account, and seasonal sales fluctuations can be red fillment, and the like) in the supply chain mitigates risk for flags. Paypal conducts no external credit checks on loan the more opaque SMEs. Most digital SME lenders are actively applicants, and approves and issues funding usually within gathering social media data and experimenting with it. For minutes. the most part, they presently rely on it only sparingly. In this context, they seek to supplement more predictive and 4.2.3 INDIA electronically verifiable cash flow data until they can prove which social media data is consistently and reliably predic- India Stack is a term used to describe a powerful public tive for SME lending purposes. digital identity, contracts, documentation and payments in- frastructure. It was created by the government in reforms Unlike China, it should be noted that e-commerce and begun in 2010, and is ushering in a new age of secure, fric- payments platforms and the emerging Payments Banks in tionless, low-cost digital financial and business transactions. India are not allowed to lend directly to SMEs. As a result, The India Stack comprises Aadhaar (biometric identity), they have actively embraced partnerships with digital SME E-KYC (sharing identity information with consent), E-Sign lenders (to-date, all SME balance sheet lenders) to provide (digital document signing), Digital Locker (secure govern- financing to their SME sellers (and buyers) to build loyal- ment document issuance, storage, and sharing), and Uni- ty and sales. Some platforms also are starting to earn new fied Payment Interface (UPI), a payments architecture that revenues from sharing SME data with lenders used in the enables universal electronic payments. In addition, in a move lending decisions. These partnerships also tend to be good to curtail the shadow economy and reduce corruption, the cultural fits because both industries have digital bases and government announced the demonetization of all RS 500 embrace advanced technology, data-driven business deci- (about US$7.70) and RS 1,000 (about US$15 currency notes) sions, advanced analytics, and fast-paced implementation on November 8, 2016. While controversial and disruptive, and growth. the move has already led to a rapid uptake in digital payment transactions and acceptance points across India. On India’s B2C and B2B e-commerce fronts Amazon India competes primarily with home-grown e-commerce com- India has a young, but fast-growing start-up digital SME pany Flipkart (which acquired rival eBay India in April 2017 lending industry. It is using these data sources and their new- and is rapidly closing in on a deal to acquire or merge with ly accessible digital transaction flows to circumvent previous financially-troubled rival Snapdeal). This is occurring even as barriers to SME credit in India. These include underreported other global companies, such as China’s Alibaba (through income, poor accounting practices and business acumen, its investments in Paytm eCommerce), China’s Tencent 18 ALTERNATIVE DATA TRANSFORMING SME FINANCE (through its investment in Flipkart), and Japan’s Rakuten, are New fintech platforms are providing novel ways of utilizing readying themselves to stake a claim in the Indian market. data to drive down the costs and improve the speed at which With just a fraction of the 30 million SMEs captured, howev- credit can be made available. Apart from analyzing tradition- er, this market remains wide open to all competitors. al credit scores of SME owners, these platforms are able to analyze business cashflow — especially for those SMEs that Alibaba entered India through the SMILE B2B platform on are able to utilize online accounting systems and/or able to Alibaba.com for India’s SMEs, and through a 62 percent access bank account information. This also includes data on ownership stake in the newly formed Paytm eCommerce. It the strength (credit-worthiness) of the customers (buyers). appears likely that Alibaba will refocus its India ecommerce growth efforts on the Paytm assets, and apply alternative Some of the alternative data used by companies like Kick- data lessons from the Chinese to the Indian market. In July further include: average margin; annual revenue versus 2016, Paytm launched small loans for its e-commerce and financing amount requested; the Alexa global rank on traffic offline sellers, with Capital Float lending at launch and other flowing to the business websites; third party reviews of the lenders, including Aditya Birla Finance and Capital First, in business and/or their products and services; the percentage the pipeline as potential partners. The loans are based on the of financing amount covered by existing purchase orders; Paytm merchant transaction histories on its platform, plus and the social network outreach of the business on Face- additional proprietary data contributed by the lenders. book, LinkedIn, Twitter, and Instagram. B2B e-commerce marketplace players such as AmazonBu- Electronic invoicing providers such as Tungsten, Basware siness.in, Alibaba.com’s SMILE platform, Industrybuying. and Tradeshift allow SME suppliers and their business cus- com, Power2SME, Tolexo, Bizongo, Moglix, TradeIndia. tomers to exchange electronic invoices — without the need com, and ofBusiness.com are expanding rapidly in India. for supporting paper trails. While suppliers can utilize any Capital Float is one of the most active digital SME lenders. of their existing online invoicing systems, even SMEs with- It now has partnerships with almost all the major B2B and out billing systems can utilize the third-party fintech’s web B2C e-commerce platforms, followed by Lendingkart. portals to create and track their own e-invoices online. They NeoGrowth lends to sellers on Flipkart while Capital First can do so without needing to purchase additional software. provides AmazonBusiness.in SME loans. These new “Intelligently linked supply chains” providers like Basware provide both lenders and suppliers themselves with 4.3 SUPPLY CHAIN FINANCE PLATFORMS the ability to more closely track transactions by utilizing on- line dashboards and tools that enable discount financing Invoice financing and supply chain and trade credit have arrangements. become important methods to facilitate access to credit for many SMEs that operate in the B2B sector. Invoice and trade Online B2B supply chain payments from accounts receiv- credit occurs when SME suppliers selling to larger businesses able (sending invoices) to accounts payable (receiving in- defer collection for a period of time after delivery. Regarding voices) and cloud-based technologies like Traxpay allow supply chain finance (SCF), suppliers have the ability to fi- access to improved financing solutions. These new solu- nance their receivables at a discount. In the U.K., it is estimat- tions avoid many of the challenges associated with more ed that 80 percent of B2B transactions are undertaken using static-based payment models offered by banks. With tra- credit. Globally, the estimate is that almost US$74 trillion of ditional static-based payment models, it is difficult for the SME business is conducted on these types of credit terms.51 “who, what, when, where, why, or how” of the transaction to change without significant costs and delays. When any of While these approaches to finance have been around for a these variables change in the course of a transaction, pay- long time, the difference today is that far more SMEs can ment and financing options are delayed and often require be served. They now qualify for these products because manual intervention. With cloud-based solutions that offer a the new digital data platforms can extend these services to 360-degree view of the business payments along the supply smaller firms on a more cost-effective basis. This is happen- chain, flexibility can be achieved by allowing variables in the ing more and more in countries and markets where SMEs are transaction to change. This facilitates supply chain financ- able to move their accounting and B2B relationships online ing in real-time as it directly connects into and monitors the and to digital platforms. These new models link the various transaction through an adaptive rules-based engine, which transaction parties (buyer, seller, and financing provider) to also reduces the risk of lenders as they can trigger financing improve business efficiency and lower financing costs. This to coincide with a transaction. includes utilizing electronic channels to improve collabora- tion along the supply chains and achieving comparative cost advantages. G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 19 Several fintech companies, such as Remitia and ApexPeak, Other companies, such as Wave — which is partnering with are also currently working on using statistical modeling to Barclay’s, are creating a peer-to-peer decentralized network approve invoice financing. Companies such as Remitia are that connects all carriers, banks, forwarders, traders and oth- quantifying the risk of paying an invoice on receipt before er parties involved in the international trading supply chain. the usual approval process. This new statistical invoice mod- Using decentralized technologies, all communication be- eling allows for a deeper analysis into accounts payable data. tween these parties will be direct and will not pass through a It helps to predict the type of approval risk an invoice sub- specific central entity. This is making financing supply chains mitted from an existing or new supplier may bring. Although and trade financing easier than the more paper-based the pricing of payment risk for unapproved invoices is not approaches used in the past.54 new, the notion of taking historical payment and invoice files to quantify and group different risks to arrive at the probabil- In its first application, Wave tackled the shipping industry’s ity that a submitted invoice will ultimately be approved and arcane “bill of lading” document, necessitated by the lack paid without any modifications is new. It takes the concept of trust in international trade documentation systems and into the account by using alternative data from a buy-side virtually changed since the 17th century. It is a receipt given systems perspective.52 to the sender of the goods that provides proof of shipment and ownership of the goods while they are shipped by sea. What is important about these models is that by combin- Once the goods arrive at their destination, the sender ships ing payment analytics to accounts payable and procurement the original bill of lading, usually by overnight courier, to the data sets, there are now new opportunities for predictive recipient, who uses it to claim the merchandise. Each in- modeling that can enhance invoice financing. As Jason ternational shipment also involves a host of other players – Busch, one of the founders of Remitia points out: from banks that loan the money to pay for the goods and insurance companies that are liable as the merchandise plies Imagine, for example, being able to estimate the prob- the seas (each with a lien on the shipment) to government ability of a purchase order being fulfilled exactly as customs inspectors, who need to check the goods and specified even before issuance, or how small, system- make sure they match up on the documentation. The more recommended changes to a purchase order could result in hands in the pie, the greater the possibility of confusion, loss a better outcome. Or think about being able to dynamically and fraud, the result of which can be never-ending lawsuits sub-out [remove] different payment mechanisms without as each party blames the other for negligence. The Wave a user even being aware of it, such as masking a innovation, based on blockchain technology, ensures there p-card-type payment model with a proprietary one that is no possibility of fraud or falsified documents.55 captures an even larger rebate, but goes through an invoice consolidator or prime partner to circumvent the card New developments in the area of linking blockchain tech- companies.53 nologies56 to supply chain finance are also showing some promise. Blockchain technology is now being applied Some companies such as ApexPeak are also able to utilize to things like smart contracts that show the potential to alternative data to facilitate invoice financing by conducting drastically reduce the need for more traditional Letters background checks on suppliers and buyers. This allows for of Credit.57 For example, some companies are utilizing the creation of some new models to better manage the risks ledger-based blockchain technology for invoicing and pay- of financing fake transactions, as well as using a data-driven ments, which also helps to facilitate access to finance. Using credit scoring engine to assess success or failure rates. blockchain technology allows buyers to approve invoices once they are delivered. Similar to the other supply chain- India-based Kinara Capital and Tanzania-based GO Finance financing models that rely on dynamic transactional data, analyze sales and delivery data for SMEs inside vertical new blockchain technologies can facilitate payment to fi- supply chains, alternative decision data that is particu- nanciers of an invoice — even if there are multiple parties larly valuable where SME credit data is otherwise scarce benefiting, including all parties from the buyer to the sup- or non-existent. Such companies obtain this data from plier as well as the lender.58 This is making financing supply partnerships and technical integrations with fast-moving chains and trade financing easier than the more paper-based consumer goods companies, vertical supply chains (for approaches of the past.59 example, Airtel, and Coca Cola); and network partners (for example, retail chain Mother Earth for artisan, cooperative, and fair trade suppliers). These partnerships have made it possible to digitally assess credit to lower acquisition costs and provide flexible terms at below market rates. 20 ALTERNATIVE DATA TRANSFORMING SME FINANCE 4.4 MOBILE DATA-BASED LENDING The telecommunications use history of potential new MODELS M-Shwari borrowers is assessed against these scorecard var- iables and a score is assigned. The cumulative score of all Mobile data-based lending models are characterized by the the variables enables CBA to make an informed choice about offering of instant small mobile loans using credit scores which new clients to provide an initial loan to and which to based on mobile transactions, mobile e-money usage, pass on. Repeat loans are then also based on past repayment mobile e-money linked savings history, as well as prior history. Although most of these loans are for very short-term credit history data. In addition, new third-party mobile-based needs, as loan sizes expand, SMEs are also making use of lenders are using data from apps running on smartphones in these products — especially the larger loan sizes now offered alternative credit scoring models. These data sources can in- by competitors like the Kenya Commercial Bank (KCB). clude SMS messages, emails, the number of people you call or text in a day, geo-based location, social network usage, Similar to the partnership between CBA and Safaricom, and retail receipts. KCB also partnered with Safaricom to offer a rival product called the KCB M-Pesa account. Like M-Shwari, the KCB M- Even obscure variables can bear on a decision to extend PESA account is a virtual mobile-based bank account that is credit. These variables can include: how frequently a user offered to M-PESA registered customers. It allows them to recharges the phone’s battery; how many incoming text save funds in the bank and earn interest, as well as to bor- messages a client receives; how many miles the client trav- row micro loans.61 Although the features of the two linked els in a given day; whether a client gambles; and even how bank services are similar, the KCB product offers much a client enters contacts into their phone (for instance, the larger amounts, and lower loan interest rates and longer decision to add a contact’s last name correlates with cred- terms. These terms make it more viable for micro and SMEs itworthiness.)60 with terms of up 180 days —as opposed to 30-60 days for M-Shwari — and for amounts of US$10,000). Most of the early lenders in this category have been the result of partnerships between mobile e-money operators owned One of the interesting features of the KCB M-Pesa loan is by telecommunications companies and banks. At least that it offers an auto-debit feature. This feature first debits one bank has now entered this market. In addition, there the KCB M-Pesa account on the payment date. If there are has been a rapid expansion across parts of East Africa and insufficient funds in that account, the other KCB accounts or other countries, where mobile-enabled e-money accounts M-Pesa account will be debited. This is quite different from are expanding rapidly and being used by consumers and the M-Shwari loan where payments need to be made man- SMEs. Banks and e-money operators partner and make use ually by the borrower. Both rely on alternative data based of their collective data. Banks can not only utilize data from on mobile enabled e-money transactions and airtime usage mobile e-money providers, but also utilize their own data rates, as well as more traditional information on credit histo- from the client’s savings and credit histories. Over time, as ry of previous loans and on-time payments by customers. customers take up the use of smartphones, the data richness should enhance the ability to offer mobile data-based lending, In addition to banks, there are several third-party lenders re- including for SMEs. lying on smartphone data to develop alternative data credit scoring models. These include companies such as Branch, The first mobile-based lending partnership model that which operates in Kenya and Tanzania US-based Tala reached scale was Safaricom and the Commercial Bank of (previously InVenture), which operates in Kenya and the Phil- Africa (CBA). They launched their M-Shwari product in 2012. ippines. Tala also has plans to ex­tend operations into other Since then, this joint product offered by CBA has been lend- markets. ing to small borrowers using an algorithm based on cus- tomer use of Safaricom services. For first-time borrowers, Branch uses credit-scoring models that analyze short mes- the credit-scoring algorithm consists of a set of Safaricom’s sage service (SMS) logs, social network data, call data, GPS data related to airtime, airtime credit, usage of Safaricom’s data, and contact lists. Tala uses a similar range of data sets e-money product (M-PESA), and the length of time as a cus- to determine credit worthiness. Some specific examples in- tomer of the carrier. Each variable has differing weights and clude the fact that those clients that make regular phone scores based on its predictive power. G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 21 calls with close contacts and with their own customer bas- with a good network diversity, including more than 58 con- es increase repayment predictability by four percent. This tacts, make better borrowers.62 Tala also determined that ap- is especially relevant for micro-businesses to demonstrate plicants who organized at least 40 percent of their contacts repeated business relationships with their customers. The using both their first and last names were 16 times more like- same is true for business owners who stay near their busi- ly to repay on time because this tendency demonstrated the nesses and have consistent travel patterns. They have a six organizational skills of the borrower. percent better repayment performance. In addition, those G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 23 Digital Lenders and Bank Convergence: The Future of SME Finance Initially banks may have been blind to the potential of dig- new customers to grow.65 Developing market players have ital SME lenders, and digital SME lenders may have high- the extra challenges of reaching out to and educating less lighted their disruptive models. However, now both parties literate or technology-adept SME segments. Response rates have come to a simple conclusion: that is, that there is more are typically modest, and low approval rates can mean the strength in convergence and collaborative models. cost per successful applicant can be quite high. High acqui- sition costs can still make economic sense if the platform 5.1 ADVANTAGES AND CHALLENGES FOR gets a lot of repeat borrowers; however, high customer DIGITAL SME LENDERS churn translates into a big profitability problem. First, digital SME lenders have competitive advantages over Although alternative lenders may enjoy potentially lighter existing banks because of convenience, cost, speed, tech- regulations and compliance burdens, there are clear signs nology, and new data credit screening. Convenience — across many country markets that more regulations are all clean, simple, friction-free applications, fast decisions and but inevitable as the industry grows. Digital lenders are cur- funding — wins business for these marketplaces, even where rently operating under a plethora of regulatory and licens- rates and fees are significantly higher than banks. ing schemes. Some countries have not yet implemented any specific regulations (India); some waited until lending vol- Second, unburdened with branch networks, existing legacy/ umes reached sufficient scale or problems emerged before silo-based systems, and often lighter regulatory and com- implementing regulations (China); and still others proactive- pliance costs, it is also true that many digital lenders have ly implemented regulatory regimes early on that balance a radically lower lending costs (for example some report their supportive regulatory environment with safety and sound- lending costs are 400 plus basis points lower as a percent of ness while also promoting competition (U.K.). outstanding loans63). Bankers simply cannot close this effi- ciency gap through cost-cutting measures alone. At the same time, some fintech companies have obtained or are pursuing full banking licenses, most notably the tech- However, three items on the lending cost bar — including nology giants in China — Alibaba, Baidu, and Tencent — but the cost to acquire, the cost of credit losses, and the cost of also Tyro Payments in Australia, Solaris in Germany, and capital — dwarf the cost of maintaining a client relationship. Klarna in Sweden. Likewise, a number of all-digital or mo- Non-bank digital lenders have a significantly higher cost of bile-first, mostly consumer-focused banks have emerged capital than their bank competitors, which can leverage their in the last two to three years. Only a handful, though, are nearly free deposits for liquidity. In today’s environment, a or are planning to focus heavily on the SME segment, in- bank’s cost of funds for a line of credit will typically be in the cluding: Finland-based Holvi (acquired by BBVA in 2016); range of 50-60 basis points—a fraction of the 600-1,200 ba- Mexico-based Bankaool; Russia-based Tochka Bank; UK- sis points marginal cost for non-bank lenders.64 The capital based CivilisedBank and Tide Germany-based Penta, and markets can also seize up without notice, leaving these plat- US-based SEED. forms without the capital to meet borrower demand. Indeed, this was an issue faced by U.S. platforms in particular in the 5.2 ADVANTAGES AND CHALLENGES first half of 2016. FOR BANKS Most online alternative lenders also have a relatively high cost Existing banks have strengths that they can use as they de- of customer acquisition, as most are still building awareness. fend against digital SME lenders. First, banks have large, They must use channels such as marketing ads and incen- captive customer bases and a recognized brand. SMEs walk tive offers on online platforms, television or radio advertising, into their branches on a regular basis without prompting. direct mail, brokers, and loan aggregation portals to source Demand for or awareness of traditional bank products, 24 ALTERNATIVE DATA TRANSFORMING SME FINANCE even without the speed and agility of the new digital 5.3 COLLABORATIVE PARTNERSHIPS GAINING lenders, still exists — at least in the more developed markets. MOMENTUM Further, banks rely on their own SME customers to proac- tively come to them when they need funding. Alternatively, Many see collaboration between banks and SME digital lend- banks can pre-screen customers using their own data for of- ers as a logical step in the industry’s evolution. The combina- fers. As would be expected, the risk profile of these existing tion of fintech and bank advantages has the potential to cre- SME customers of banks skews more in a positive direction. ate truly sustainable and scalable competitive advantages for In contrast, direct lending that targets new SME clients tends both partners along with a significantly lower cost structure to attract a larger number of marginal, less credit-worthy for digital SME finance. But, it is not a foregone conclusion. and even a certain number of fraudulent applicants, thereby Successfully navigating partnerships between fintechs and driving up acquisition costs. banks is challenging: the two cultures do not always mesh, it can be difficult to overcome inherent competitive conflicts, Second, banks have extraordinarily valuable internal SME and regulatory due diligence can be onerous for fintechs data and all of it is virtually free, beginning with the powerful that are not prepared. SME free cash flow data, as evidenced by business current and POS merchant accounts. However, unlike digital SME Nevertheless, the options for how banks and SME digital lenders which use this data as a staple, most traditional banks lenders can work together have increased significantly and ignore it for SME lending decisions. Although digital lending will likely continue to expand as the stakeholders become partnerships are starting to make the banks’ own data availa- more creative. They range along a spectrum of “light touch” ble for online SME lending decisions, banks which make the to “deep-touch” linkages between the partners. The lighter leap to build their own digital SME lending platforms could touch end of the spectrum includes one-way or two-way instead find, underwrite, and book loans at near-zero cost—a customer referrals and offering bank loans on SME loan bro- massive advantage over non-bank lenders. ker marketplaces,66 as well as bank direct investment in SME digital lender loans for yield. Deeper touch options include Third, banks’ low cost of funds from customer deposits place SME data-rich information exchanges, deep strategic and them at a unique advantage over non-bank SME lenders. technology integrations, new distribution channels, equity stakes in digital lenders, joint ventures, acquisitions, and in- Fourth, although startups may be agile, banks can out- novation centers, incubators, or accelerators (Figure 10). perform them when it comes to managing massive amounts of money. Banks are stable and their technology systems 5.3.1 Light Touch Partnership Models move very large amounts of money around the globe via transfer systems like Automated Clearing House (ACH) and Lighter touch agreements keep current operations of the Swift. Banks also have decades of experience in lending, partners at relative arms lengths. First, banks can refer SME navigating the regulatory maze, and compliance — in good customers they have turned down for a loan to SME digital times and in bad. Box 1: Why Banks and SME Fintech Lenders are becoming more friends than foes Advantages for Banks Advantages for Fintechs Captive, large customer base/ Superior customer experience positive selection in applicant mix Clean, simple, and friction-free applications Brand More credit data sources Distribution coverage Advanced risk models/risk-based pricing Extraordinarily valuable, “free” internal data Underwriting costs Low-cost, stable source of funds Potentially lighter regulation (but high future Regulatory certainty (mostly) uncertainty, more regulation) G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 25 Figure 10: How Banks Can Compete with Online Lenders © 2017 Karen G. Mills and Brayden McCarthy. Reprinted by permission. Retrieved on April 27, 2017 from: Karen Mills and Brayden McCarthy (April 26, 2017). “How Banks Can Compete Against an Army of Fintech Startups.” https://hbr.org/2017/04/how-banks-can-compete-against-an-army- of-fintech-startups Karen Mills is a senior fellow with Harvard Business School and Harvard Kennedy School and served as Administrator of the U.S. Small Business Administration (SBA) from 2009 to 2013. Brayden McCarthy is vice president of strategy for Fundera, and served previously as senior economic policy advisor at the White House and the SBA. lenders in exchange for a referral fee. While these agree- Second, creative cross-referral relationships can be an ments can work well, they can be frustrating for both parties appealing approach. In Singapore-based DBS Bank’s as well as SMEs if the bank does not generate enough quality cross-referral agreements with Singapore-based P2P SME referrals and the SME digital lender funds relatively few of lending platforms Funding Societies and MoolahSense, the bank’s turndowns. for example, the bank refers first-time SME borrowers to the platforms and the platforms refer SME borrowers who In November 2016, the UK government made it mandato- have completed two successful loans back to DBS to access ry for nine of the nation’s top banks to refer their rejected larger loans and/or for additional financial products. In an- SME loan applicants to SME digital lenders. The government other example, UK-based Santander proactively refers small initially designated three SME loan broker marketplaces to business customers looking for a loan to Funding Circle on connect unsuccessful SMEs with alternative SME digital its website and in letters to customers, while Funding Circle lenders:  Funding Options, Funding Xchange and Bizfitech. signposts borrowers to Santander for banking relationship, These platforms each include a panel of alternative lenders cash management, and other banking services. that offer a variety of SME financing options. Tracking what impact this mandate has on SME credit access and satisfac- A third option is for the bank to offer its SME loan prod- tion in the UK will provide important industry insights for ucts on SME loan broker marketplaces or e-commerce sites. bank referral relationships. These partnerships let banks explore digital distribution of 26 ALTERNATIVE DATA TRANSFORMING SME FINANCE SME loan products. Banks gain the opportunity to acquire 5.3.2 Deep Touch Partnership Models pre-vetted SMEs desiring loans that match banks’ desired SME borrower profiles established with the portal while the Deeper touch partnerships have emerged in the last two broker/site provider earns referral and other fees. Through years and can have a wide variety of structures. In general, these portals, banks may also acquire new SME customers the SME digital lender provides access to and integrates all or from outside their existing footprint. If the bank chooses to parts of its proprietary technology for applications, pricing, offer a suite of their SME lending products, the more robust underwriting, servicing, and/or monitoring; the bank pro- product offerings may give it an advantage over single- vides the credit policies and underwriting criteria; and which product non-bank lenders on the platform. partner funds the loans varies. The SME digital lender brings the many external sources of alternative and traditional data A fourth option involves banks that provide lines of credit to it taps in the bank partner’s market footprint (and adds new fund loans or buy loans originated on SME digital lenders’ country-or-region specific data partners as needed). It also platforms. For example, in the US, JPMorgan Chase, Bank uses its technology to directly and more efficiently access of America, and SunTrust have been active in buying loan the bank’s own SME customer data than the bank can do on assets from leading SME digital lenders. To free up capital its own. A critical question to answer in this type of structure for more loans and remove risk from their balance sheets, is whether (and the degree to which) the bank retains its own digital lenders securitize a pool of SME loans originated on underwriting and SME scoring or relies upon the SME digital their platforms to sell to banks and institutional investors.67 lender’s scoring models and algorithms.68 Banks then purchase these securities as a way to diversify investments, gain a new source of balance sheet growth Several examples illustrate how these deals can work. Draw- with attractive yields, and put their excess deposits to work. ing on a recent strategic shift to focus more on providing They also decide on the types of assets they are willing to its technology to banks and other lenders under its new buy (such as high, medium, or low risk). Banks that buy these “lending-as-a-service” (LaaS) platform, OnDeck launched securities must typically develop specialized expertise to un- a partnership with JPMorgan Chase in the United States derstand the lender’s overall and securitized loan pool risk in April 2016. Under the deal, Chase white-labels OnDeck profiles as well as how its loans might fare in an economic technology to make faster decisions for online SME loan turndown. As a rule, banks therefore generally only want to applications while funding and shouldering the risk of the work with established SME digital lenders that provide full loans going bad. In mid-2016, Singapore-based DBS Bank transparency and sufficient loan performance data and his- partnered with Hong Kong-based AMP Credit Technologies tory to conduct a comprehensive risk/reward analysis. (AMP) to use its “software-as-a-service” (SaaS) technology to launch the white-labeled “DBS mLoan.” They designed While providing a line of credit or buying loans offers finan- the loan product for SMEs that accept card payments (such cial benefits for banks, they offer no benefit to a bank’s SME as retailers, restaurants and e-commerce merchants) and customers or brand. To address this deficit, some banks in- have been in business for at least six months. AMP’s technol- stead use an SME digital lender’s proprietary analytics to un- ogy combines credit modeling with daily cash flow data in cover additional SME loan opportunities in the bank’s current order to enable lenders to offer unsecured SME loan prod- SME portfolio. The SME digital lender then directly offers and ucts. The technology also direct-debits repayments from funds the loan for qualifying SMEs, and pays a referral fee to the borrower’s core business operating account (minimizing the bank. Usually banks can then also optionally buy back friction and risk) as fixed, equal installments every banking loans that meet their portfolio criteria. business day.  Some governments also participate in buying SME digital Partnership agreement structures can also facilitate SME lenders’ loans, which also helps build credibility and aware- digital lenders’ global expansions and/or include equity ness for an emerging SME digital lending sector as well as deals. Kabbage, which began licensing its Kabbage Platform decent investment returns. This was the case for UK-govern- to other lenders In March 2015, has implemented its tech- ment-owned British Business Bank (BBB), which has made nology at ING in Spain, Santander in the U.K., and Scotia- funding commitments totaling about US$220 million to bank in Canada and Mexico. The three banks plan to roll nine carefully vetted UK SME digital lenders since 2012. The out the technology across the rest of their respective global program is set up to fund a gradually declining percentage footprints. Each bank also took an equity stake in Kabbage. of each loan made as the lender’s loan volume and market Kabbage’s technology allows the banks to provide fully penetration increases. As a result, the BBB commitments to automated SME applications, underwriting, servicing, and date have (or will) support new lending volume to UK SMEs monitoring of loans, while the banks set the underwriting of almost US$1 billion. criteria. Funding for the SME loans differ by bank; for exam- ple, ING and Kabbage both fund the loans, while Scotiabank G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 27 issues all of the loans on its balance sheet but gives Kabbage and will later expand across Asia. Executives describe the an option to buy back a portion of the loans. joint venture as “building symbiotic relationships with Kore- an and international startup companies, financial firms, and Some deals instead co-brand the SME digital platform with other FinTech industry members to create a robust Fintech the bank, offering loans, “powered by” the SME digital lend- ecosystem.”69 er’s technology and underwriting. US-based Regions Bank entered into a partnership with US-based SME balance sheet Establishing innovation centers, incubators, or accelerators lender Fundation in October 2015. The platform is co-brand- and selecting SME fintech innovators to support is another ed with Regions with the loans “Powered by Fundation,” and strategy banks are pursuing that is starting to yield tangi- Regions’ suite of SME products sits alongside the Fundation ble results. For example, in June 2016, Netherlands-based loan option. Fundation’s technology provides digital tools ABN Amro and HighTechXL  teamed up  to build an “open like online applications, real-time third party data aggrega- innovation ecosystem” they dubbed “Econic” to accelerate tion, and data-intensive proprietary decision techniques to innovative ideas coming from both outside and inside the predict credit risk and appropriately price loans. SMEs apply bank. Netherlands-based invoice management firm Invoic- for a loan online through Fundation, but borrowers can also eSharing was one of the first six fintech startups accepted elect to have a call with a Region’s banker to complete the into the program. Just eight months later, the bank and In- application. Depending on the loan type and amount, either voiceSharing collaborated to launch a comprehensive solu- Fundation or Regions funds the loan.  tion that provides SMEs with 24/7 insight into their accounts to estimate their working capital needs well in advance. An Canadian-based SME digital lender Thinking Capital struck accounting robot tool reads and checks the invoices, gen- a co-branded “Rapid Financing” SME lending platform and erates journal entries, and exports the invoices to the entre- cross-referral partnership with Canadian Imperial Bank of preneurs’ accounts and accounting system. The robot also Commerce (CIBC) in November 2015. CIBC offers their SME compares invoices with historic data from industry partners, customers a new set of Rapid Financing loans alongside using accountancy data based on the preceding three years. other CIBC SME financing products. SMEs can apply for a SME clients save time and money, while ABM Amro builds loan in under 10 minutes, enjoy instant decisions, and faster SME loan volume. funding. CIBC also offers incentives to Thinking Capital loan customers to move their business banking to CIBC. As banks and digital lenders become increasingly accus- tomed to working together, opportunities for strategic ac- Other partnerships jointly develop lending products target- quisitions will come into view. These will bring a new set ing specific SME customer segments. In late 2016, Capital of issues such as valuation differences between banks and Float partnered with India-based IDFC Bank to provide dig- leading SME digital lenders and the integration of very dif- ital lending that will focus on SME borrowers who have no ferent cultures. access to bank credit, with limited or no documentation and without existing credit history. IDFC Bank will gain access to To date, only two bank acquisitions of SME fintechs of note Capital Float’s digital network of borrowers, thereby ena- have taken place. In March 2016, Spain-based BBVA ac- bling it to diversify its portfolio of small ticket loans and grow quired Finland-based Holvi, a startup that specializes in pro- its customer base. Capital Float, in turn, can leverage IDFC viding online current accounts and related services for SMEs, Bank’s balance sheet, product innovation, and customiza- entrepreneurs, and freelancers. The Holvi platform allows tion of banking products for this segment of borrowers. BBVA to access SME customers at a low cost of acquisition with high cross-selling opportunities such as lending and Distribution partnerships and joint ventures offer other foreign exchange. In July 2016, France-based Group BPCE promising paths for regional or global expansions. For exam- (the parent company of two major Banque Populaire and ple, OnDeck entered the Australian market with Common- Caisse d’Epargne cooperative banking networks) acquired wealth Bank of Australia (CBA) and Australian-based online Germany-based Fidor, one the earliest “fintech” banks. Fidor accounting software provider MYOB as distribution partners offers digital banking to SMEs and also offers its banking li- (MYOB also took a 30 percent equity stake in OnDeck Aus- cense and proprietary open API digital banking technology tralia). In February 2016, China-based Dianrong established platform to other companies. Group BPCE sees the acquisi- a 50:50 joint venture with South Korea-based conglomerate tion as a key step in accelerating the company’s digital trans- Hanwha Group to offer loans and other financial services in formation while Fidor sees it as enabling a strong interna- South Korea. Dianrong will bring technologies while Hanwha tional expansion, more technology innovation, and a bigger takes on marketing. The joint venture will begin by launching presence in Europe. an open P2P lending+ marketplace in South Korea in 2017, 28 ALTERNATIVE DATA TRANSFORMING SME FINANCE In summary, while deeper partnerships can provide banks banks. The new technology being used by the bank helps to with a faster time to market for SME digital lending than enhance and fast track customer on-boarding, account building a platform from scratch, they require top executive opening, and loan approvals. The integrated system provides support, clearly aligned expectations and goals, and signif- a comprehensive view of the customer’s banking relation- icantly more bank commitment and resources than lighter ship, across business and personal transactions, along with touch options. The due diligence process must cover a myr- all records from all accounts with the bank. This eliminates iad of issues mandated by regulators and bank prudence, the need to re-enter data or provide additional information such as technology integration, risk models and manage- for loan processing or new accounts. It also provides access ment, compliance, data security and privacy, fair lending, to information that is often not available in traditional banks and reputation risk, and can be lengthy.70 The partners also where customer information is held in different systems or need time to become comfortable with and trust each oth- siloed in different departments. er. That said, if the partners take a long-term view of the relationship, these partnerships have the potential to provide CivilisedBank and the new breed of digital bank are examples transformational value and growth opportunities for both. of not needing to separate customer relationship manage- ment (CRM) systems to bridge between their different mar- 5.4 THE RISE OF THE DIGITAL SME BANK/ ket segments. Consumer and SME finance are set up in their LENDER new, core system as a unified database, with inbuilt CRM ca- pability. By contrast, older banks often have separate systems Building a data-driven, customer-centered relationship with that must be bridged by expensive, complex CRM add-ons. SMEs constitutes a new frontier for many banks. Several new and a few existing banks are directly developing their own Launched in October 2015, US-based startup Clearbanc tar- in-house SME digital lending systems along with opening up gets freelance SMEs like Airbnb providers and Uber drivers. It their application programming interfaces (APIs) to third-par- provides credit using a “cash advance” approach that relies ty data and service providers. Some large traditional banks on analyzing alternative data.72 For example, drivers give the are also opening stand-alone all-digital banks. bank access to their Uber accounts to verify the hours they work and how much they earn on average. This allows Uber New, or neo, banks71 often have a unique advantage over drivers who are paid weekly to receive their daily earnings larger and more established banks because they do not have in advance through a Visa debit card. Clearbanc is also of- the legacy systems or heavy operating costs that tradition- fering similar products to Lyft drivers. In addition, plans are al banks have built up over decades and which are harder underway to use online accounting systems, such as Intuit’s to change. These new players are also able to more quick- Self-Employed Solutions accounting software and Quick- ly adapt and utilize alternative data in making their credit Books software for freelancers. decisions. Several existing large and established bank players have Some of these new banks have built a “banking-as-a-plat- launched their own digital data-based SME lending models, form” (BaaP) model. This allows for improved cooperation including some that are supported by alternative credit scor- with alternative and third-party fintech models. It also allows ing analytics specialists that have emerged in the last sever- for modular banking platforms that cater to mobile-first al years. Kenya’s Equity Bank and Airtel launched Equitel, a customers, real-time banking, as well as peer-to-peer and Mobile Virtual Network Operator (MVNO), offering loans of crowd functionalities within the framework of an open up to US$30,000. With loan terms as long as one year and API-based infrastructure. In addition, by cooperating with interest rates as low as 1.5 percent (a flat monthly rate), Eq- third-party providers, the bank provides a platform for finan- uitel’s mobile loan products are the longest term and lowest cial innovation in much the same way that Apple acts as a cost mobile-enabled loans in Kenya. Although Equitel does platform for developers through its App Store. With peer-to- not have the rich mobile data history from a mobile network peer lending and cooperation with marketplace lenders as operator, it does have an extensive history with its own cli- well as open APIs to support trade and supply chain finance, ents, which it uses to provide a credit score and determine these open bank platforms should continue to expand the various risk-rated interest rates to be charged.73 financing options for SMEs. India-based The State Bank of India (SBI), has also made New online-only start up CivilisedBank in the U.K. focuses use of alternative data as well as partnerships with innova- exclusively on SMEs. CivilisedBank is using a special back- tive business networks, such as Uber and its drivers, as well end system which allows for integrated banking, payments as with e-commerce marketplace providers such as Snap- and improving risk compliance while also focusing on deal, Flipkart, Amazon, ShopClues, and Paytm to provide alternative lending for SMEs in collaboration with other local micro and SME financing. In February 2016, the bank G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 29 launched the  SBI e-Smart SME, a collateral-free working aggregators, solar companies and other digital product and capital loan offering for sellers on e-commerce platforms. service providers. Sellers apply for the loan online and receive an immediate answer with one click. The product relies on proprietary The company’s scoring technology looks at four kinds of e-commerce platform data about the seller’s sales and other data: demographic, geographic, financial, and social. The data, as well as surrogate information from the public do- first two help verify identity and provide context. Financial main to assess the seller’s credit worthiness for loans. For data provides a pattern of usage, not only mobile money, Uber, the partnership allows SBI to access data from the but also regular top-up transactions that offer a window into driver’s Uber history. It simplifies the documentation require- consumption patterns. First Access analyzes existing client ments, thereby eliminating traditional financial statements data from FINCA’s operations, as well as subscriber data from like income tax returns. local mobile network operators (MNOs) to establish credit scores for clients to secure small loans to build businesses Use of open APIs and support for integrating bank data with or support emerging personal needs. As First Access collects online accounting platforms are also strengthening the use data, it will also continuously recalibrate its dynamic FINCA of digital data to better enable banks to address SME lending. algorithms using machine learning techniques and hands- U.S.-based Wells Fargo, which launched its online FastFlex on collaboration with its data scientists, enabling FINCA to Small Business Loan in May 2016, created an API so that SMEs refine its own product offerings and improve credit quality. can have their bank account data uploaded directly into the Through its partnership with FINCA, First Access is expanding accounting software provided by Xero. By digitally connect- its current Africa footprint from Tanzania and Kenya to Zam- ing with the bank, SME customers see their real-time, up-to- bia, Democratic Republic of Congo, Uganda, Malawi and Ni- date cash flow each morning on Xero and can receive pay- geria, where FINCA maintains subsidiaries. ments faster. This also facilitates credit decision-making and allows the bank to better analyze the business as a whole. Other banks have opted to launch all-digital SME banks. DBS, which has been undergoing a 3-year digital transfor- Similarly, New Zealand-based Heartland Bank is a small mation to become a digital banking platform with open APIs, challenger bank. It is growing by targeting niches that the launched its mobile-only DBS Digibank for consumers in big banks overlook, including SMEs and rural lending. In April India in April of 2016. Taking advantage of the existing in- 2016, it launched a fully-automated Open for Business unse- frastructure in India - mobile phone and internet penetra- cured loan for SMEs. In two to three minutes, SMEs can apply tion, Aadhaar card-related and PAN card-related information for up to US$ 35,000 by answering just six questions regard- (gathered and verified by the government) - DBS’s Digibank ing their identity, amount to borrow, income, whether they is completely branchless and paperless. DBS has now gone have a mortgage, and whether they use Xero or MYOB ac- one step further and is in the process of launching its digital counting software to run their businesses (indicating to the SME platform in India as well. DBS’s digital banking platform bank that the business is well-run). Heartland’s computers is allowing it to expand internationally rapidly, with Indonesia “talk” to credit rating agency Veda to establish whether the targeted next for its Digibank platform rollout. SME owner has a clean credit file and whether the business has the capacity to manage loan repayments. The applica- The recent emergence of full service cloud-based SME lend- tion process takes just two to three minutes, and borrowers ing technology platforms offered by a growing number of receive an immediate decision communicated online. emerging banking technology firms have made in-house SME digital lending and starting a new all-digital SME plat- In May 2016, Washington D.C.-based the FINCA global mi- form bank more viable and attractive. They can significantly crofinance network, which provides nearly 2 million people reduce the time, costs, and complexity needed to bring these in 23 countries with financial services, including micro and capabilities to market. Despite these benefits, however, these small business loans and credit lines, launched a collabo- partnerships cannot replicate the multi-year time spent by ration with US-based First Access, whose technology pre- the older, more established digital SME lenders proving and dicts the credit risk of borrowers in informal markets. The improving their SME credit scoring model sophistication and partnership creates a sophisticated alternative credit scoring predictive reliability using alternative data that comes from approach to improve FINCA’s outreach to excluded pop- lending experience. Banks still need to go through the same ulations and apply risk-adjusted pricing across the whole iterative scoring improvement process as they bring in new spectrum of FINCA’s borrowers in Africa. The First Access alternative data sets before they can scale their SME lending Enterprise Scoring platform captures data from loan appli- significantly. They also need to acquire and have the decision cations, core-banking software, credit bureaus, smartphones scientists and technology staff capable of managing both and feature phones, unique commercial partnerships with structured and unstructured alternative data sets. mobile network operators, mobile money platforms, data 30 ALTERNATIVE DATA TRANSFORMING SME FINANCE Paradoxically, these banking technology firms also appear offer, in this case, an SME loan that best suits the needs of the to be having an impact on banks’ appetite for “deep-touch” SME borrower. Importantly too, there is also a much wider but more complicated SME digital lender-bank partnerships range of non-financial companies willing to become users in favor of in-house solutions which let a bank control the of these banking platforms in order to leverage the bank’s entire lending process.74 In essence, the “bank-cloud lending SME customer data in a way that is profitable for every party. technology platform” combination is fast becoming a new competitor category for both SME digital lenders and bank- As noted, there are a range of variants of more collaborative SME digital lender partnerships. banking models that are rapidly emerging, and will contin- ue to emerge. Advanced data mining and analytics on an Globally, entrepreneurs and some traditional banks are cre- ever-expanding set of new data sources as well as SMEs’ ating SME banks and SME lending that embrace a digital-first growing digital footprints underpin these collaborations. It strategy. The broader trend is moving the industry toward is a near certainty that fintech and banking innovators will creating a banking ‘platform’ and opening its API up to oth- continue to identify more new sources and types of alterna- er fintechs, third party developers, third party lending and tive SME digital data that can reliably be used as SME credit other financial services providers, and even other banks so data (beyond what we have covered in this report). In turn, they can combine the bank’s data with their own to build these innovators and innovations will lead to continuing new products on top of this platform. Through the API, the SME financing options and access expansion. platform bank can consult all of its third-party providers and G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 31 Alternative Data: Policy Issues and Challenges As noted, the rise of alternative data-based lending has credit-approval process. Given the lack of privacy protec- opened new and innovative approaches for financing SMEs. tions for SMEs that utilize these new lenders and/or data ag- At the same time, these changes raise a number of policy gregators who act as data-brokers, there will be a need for issues that stakeholders are only starting to understand. The additional regulation to ensure that SMEs are protected.  recently issued G20 High-Level Principles for Digital Finan- cial Inclusion75 provide guidance in a number of areas that Although a lot of analysis and recommendations about alter- can assist in developing approaches specific to alternative native lenders and consumers have been made, some —but data-based lending. not all—rules may apply to SMEs. These rules would most likely focus on:  This section identifies various policy theme areas that have • Ensuring “ownership” and appropriate rules for the use, surfaced for stakeholder consideration and further discus- security and control of individual and SME data.77 sion, including:  • Ensuring that data information will be protected and only • Data privacy and consumer protection issues  shared with other parties, and with the express approval of • Opt-in as opposed to opt-out models  the SME borrower.  • Credit information sharing  • Developing a framework78 for all financial service provid- • Cyber security and data  ers to implement transparent, user-friendly and effective • Pricing transparency  recourse mechanisms and dispute resolution mechanisms • Balancing integrity, innovation and a competitive to address SME claims and complaints including: marketplace  — Instituting a process for correcting or deleting inaccu- rate or unsolicited information.  6.1 DATA PRIVACY AND CONSUMER — Establishing a mechanism for a clear data retention PROTECTION ISSUES  period. — Setting up appropriate SME hotlines to address With the range and types of data being collected from online questions and complaints.  digital and/or mobile footprints to social media and other — Providing for external consumer complaint depart- forms of information which are being used by alternative ments within the financial regulator or appropriate data lenders, there has been an increasing concern about government agency to also address concerns of SME privacy and the need to be more transparent about how borrowers, especially for those new issues related to data is being collected and used. Information privacy laws, alternative data lending models. rules and principles also vary widely across different jurisdic- tions.76 Data sources used for identity verification and fraud 6.2 OPT-IN MODELS AS OPPOSED TO OPT-OUT can be combined with far-reaching data profiling and tar- MODELS  geting capabilities by alternative lenders. In addition, digital and mobile marketing and underwriting decision processes Several of the mobile data lenders utilize mobile and other can operate in a non-transparent manner, using so-called data often by default or with limited and/or offer obscure “Black Box” structured and unstructured data sets and algo- opt-out features. Automated collecting of mobile and digital rithmic techniques that currently operate using proprietary data has opened up a completely untapped credit market, lending models.  especially in Kenya and other countries in Africa. However, having clear but simple opt-in models will most likely be re- Although these lending platforms have provided access to quired in the near future. Clearly informing customers how those borrowers formerly excluded from credit, they often their data will be collected and used, as well as allowing these fail to provide details as to how they gather and use data informed customers to opt-in as opposed to opt-out, should for identifying prospects, credit ratings, scoring and the be something that providers should proactively consider.  32 ALTERNATIVE DATA TRANSFORMING SME FINANCE 6.3 CREDIT REPORTING SERVICE PROVIDERS  approaches for aggregating and using this data for quality lending decisions may be needed. In essence, in an environ- In many jurisdictions, alternative data lenders, especially ment where everyone can contribute data, much of it not non-bank players, are not currently required to report their traditional credit data (such as e-commerce, mobile, social, SME borrowers’ application and loan performance to credit trade, and the like), and many types of entities can use this reporting service providers. While some SME digital lend- data (not just banks or traditional non-bank lenders), mar- ers voluntarily report this information (enhancing the val- ket alternatives to traditional credit reporting services may ue of this information for the entire lending industry while be needed.80 also helping lenders attract SME borrowers with good pay- ment history and create leverage with borrowers to pay on As the G20 High Level Principles of Digital Financial Inclu- time), many more do not. These lenders most often argue sion highlight, policy makers and regulators should promote they have concerns about competitors targeting their SME the establishment and responsible use of flexible, dynamic customers acquired through expensive proprietary lending credit reporting systems modeled on best practices as out- models and acquisition costs and/or the time, costs, and lined by the International Committee on Credit Reporting complexity of regulatory compliance (such as providing SME (ICCR). These can include relevant, accurate, timely and suf- borrowers with the right and a process to dispute or cor- ficient data collected on a systematic basis from all reliable, rect credit report information). There is, however, growing appropriate and available sources, and retained for a consensus even among alternative lenders as well as poli- sufficient time period. The overall legal and regulatory cy makers that the benefits of reporting credit application, framework for credit reporting should be clear, predictable, credit and payment performance, and account closure data nondiscriminatory, proportionate, and supportive of con- to credit reporting service providers consistent with the data sumer data protection and privacy rules. reported by banks and other lenders outweigh these ar- guments. Key issues for policymakers to consider here are In addition, policy makers should encourage the use of in- whether or not to mandate (and if so when) or simply en- novative data sources in credit reporting systems such as courage more voluntary reporting and the standardization of data on utility payments, mobile airtime purchases, as well the information reported. as use of data on digital wallet or e-money accounts and e-commerce transactions.81 The ICCR provides a useful A separate, important, but much more complicated policy venue for evolving global guidance on how to handle the issue to consider is whether or not to encourage or man- growing sources and users of data described in this report, date certain alternative data reporting (such as mobile/ and the G20/GPFI should support ICCR’s continuing work in other digital payment transactions, bank account transac- this area. tions, or trading data) to credit reporting service providers. While there is general consensus that expanding the use of 6.4 CYBER SECURITY  alternative data by credit reference agencies could have a significant positive impact on access to finance,79 there are With new players and technologies being deployed, reg- many policy and process change issues to consider. These ulators and policymakers are under increasing pressure to include questions such as which data, how much data, the ensure that consumer and the public’s data are protected. inconsistent availability or quality of alternative data across In addition, many new financial service providers and/or fin- lenders or other data providers, the ideal frequency of the tech players have often not made sufficient investment in reporting (much of this data’s value lies in the fact it is re- data protection and cyber security. As such, they find them- al-time), what data governance rules apply to SMEs, data selves at increasing risk. For alternative lenders, especially furnishers, and credit reporting service providers, and many new financial players and third-party providers who support more. In addition, existing legal frameworks can impede the alternative lending approaches, basic checklists like those sharing of such data, requiring changes in laws or regulations developed by the US agency, the Financial Industry Regula- to permit data sharing. tory Authority (FINRA), should be adopted. Indeed, the complicated policy issues of sharing and col- These include:  lecting alternative data pose questions for the sustainabil- • Identifying and assessing cybersecurity threats  ity of the traditional role and business model of the credit • Protecting infrastructure and platforms from cyber reporting industry itself. As SMEs and consumers grant per- intrusions mission for lenders and other SME cloud-based service pro- • Detecting a compromise or vulnerability viders to access their growing digital footprints in exchange • Responding through a risk-based plan for a variety of value added services, including lending, new • Recovering and/or replacing lost data G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 33 • Since most alternative lenders, third party aggregators 6.6 BALANCING INTEGRITY, INNOVATION, AND and data analytic providers also store, use or electronical- A COMPETITIVE MARKETPLACE ly transmit personal identification information (names, national IDs, social security numbers, dates of birth, ad- As noted in the report by the World Bank and the Committee dresses and other key personal data) or sensitive informa- on Payments and Market Infrastructure on Payment Aspects tion (financial records, account information, tax filings), of Financial Inclusion (PAFI), specifically guiding principle 2 additional steps should also be taken to protect privacy.82 on the legal and regulatory framework, it is important to pre- serve the integrity of the financial system, while not unnec- 6.5 PRICING TRANSPARENCY  essarily inhibiting the access of individuals and businesses to well-regulated financial services. These apply in many ways Many alternative lenders do not offer, or offer unclear com- to not only payment service providers, but to other finan- parable standardized pricing policies. This is a concern not cial service providers — including non-bank financial service only for SME borrowers, but also for returns being provided providers.  In addition, the G20 High Level Principles for Dig- to P2P individual lenders and/or institutional investors. In ad- ital Financial Inclusion also covers the importance of balanc- dition, there are challenges concerning transparent pricing ing innovation and risk.84 for new products when repayment occurs as a percentage of payments flow, and not as specific sums on specific dates. The principle of competition in the financial services market- These are challenges that traditional annual percentage rate place provides clarity on the criteria that must be met to offer (APR) rules do not address.  specific types of services. It should also set consistently-ap- plied functional requirements. Regulators and policymakers Some promising practices being proposed by alternative are therefore challenged with enabling innovation, as well lenders include the launch of the Small Business Borrowers’ as competition in the marketplace. They should not hinder Bill of Rights.83 It is supported by both industry players and the entry of new types of financial service providers, new in- SME associations, including: struments and products, as well as new business models or • The Right to Transparent Pricing and Terms channels — as long as these are sufficiently safe and robust.85 • The Right to Non-Abusive Products • The Right to Responsible Underwriting • The Right to Fair Treatment from Brokers and Loan Aggregators • The Right to Inclusive Credit Access • The Right to Fair Collection Practices G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 35 Conclusions and Recommendations As noted from the examples illustrated in this paper and on their own. In addition, many are also experimenting with the tables below, the most significant and commonly used unstructured data such as social media data or website traffic data sets are still the more traditional structured data sets. to supplement structured traditional data. Lastly, there are However, these along with unstructured data sets are now many forms of collaboration with banks, supply chains and a being analyzed and accessed in new and innovative ways variety of fintechs that continue to emerge.  not previously available. This includes a more comprehen- sive analysis of data forged by a variety new partnerships — Apart from the rise of new marketplace lenders, the technol- data encompassing banking, finance, and industry, as well ogy and e-commerce giants now providing SME finance — as accounting, digital supply chain and sales information, either through partnerships or directly on their own — are at and e-commerce sales. New and alternative data includes the forefront of some of the biggest opportunities to expand mobile, online and social media data, as well as new devel- SME finance. These big technology companies have all the opments in psychometric testing.  data they need, along with the analytic prowess to crunch the numbers to assess the credit-worthiness of their SME Marketplace lending companies have stepped in to capi- sellers and buyers.  talize on the opportunity available to help meet more SME lending needs. Marketplace SME lenders use machine learn- In the US and China, these lenders have enough liquidity to ing and digital tools to extend credit to a wide array of SMEs lend on their own, as well as millions of potential SME clients. quickly and efficiently. While many of these same SMEs had That means that these organizations may become the banks’ previously been rejected by banks others are shopping for most ardent competitors in SME lending to this target market better terms and conditions or more rapid access to finance. in the near future.  These new lenders also utilize traditional credit data, es- pecially when it can be accessed in digital form. However, In India, where the emerging “big technology” companies they are discovering that electronically verifiable cash flow/ are forced to partner with lenders, the risk for banks that sales transactions and business cloud accounting software do not participate is that the other digital SME lenders may are powerfully accurate SME risk predictors of credit risk be able to keep their foothold and scale before these banks Table 1: Online B2B and Commerce Data Marketplace Tech/E- Supply Chain Mobile- Digital Lenders commerce Financing Data Based Banks Giants Lenders E-Commerce Sales and Purchasing Data x x x B2B Commerce Data x x x Supply Chain Trade Flow Data x x Supply Chain Digitized Document Data x x Logistics and Shipping Data x x Performance Data of SME Business customers x x x x 36 ALTERNATIVE DATA TRANSFORMING SME FINANCE Table 2: Banking, Finance, and Industry Data Marketplace Tech/E- Supply Mobile- Digital Lenders commerce Chain Data Based Banks Giants Financing Lenders Loan and Credit Card Data x x x x Current Account/Transactional Account Data x x x x Investment Account Data x x x Insurance Data x Online Accounting Data x x x x Online SME Billing and Payment Data x x x x Merchant POS and Sales Data x x x x SME Business Intelligence & Marketing Data x x x x Inventory Tracking Data x x x Economic and Industry Data x x x x x Table 3: Credit Bureau Data Marketplace Tech/E- Supply Mobile- Digital Lenders commerce Chain Data Based Banks Giants Financing Lenders FICO, Credit Bureau data x x x x Note: FICO= Fair, Isaac and Company. Table 4: Online Ranking and Social Data Marketplace Tech/E- Supply Mobile- Digital Lenders commerce Chain Data Based Banks Giants Financing Lenders Social Media Data x x x x x Search History x x Website History x x Online Rankings and Reviews x x x x build comparable platforms. These big technology organi- various transaction parties along the supply chain. As such, zations will have several advantages. They will continue to they provide greater insight into real-time data that can be refine the analytics they use to determine loan terms. They shared with financial service providers.  also have brand names that small businesses recognize, and many small businesses already depend on their services. Mobile data-based lending models are also increasing, espe- cially in Africa. Mobile transactions, mobile e-money usage, Digital platforms and the ability to analyze data also allow mobile e-money linked savings history, geo-based location new fintech players to facilitate invoice financing, and sup- data, tracking call usage, social media networks and mobile ply chain and trade finance for SMEs that operate in the B2B retail payment receipts are being used to provide alternative sector. These new models are providing data that link the data for digitally-native entrepreneurs.  G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 37 Table 5: Mobile Data Marketplace Tech/E- Supply Mobile- Digital Lenders commerce Chain Data Based Banks Giants Financing Lenders Mobile Call Pattern Data x x Mobile Business and Expense Data x x Mobile Recharge History x x Mobile E-Money Transactions x x Smartphone Mobile App Analysis x x Mobile Geo-Locational Data x x Table 6: Individual Data Marketplace Tech/E- Supply Mobile- Digital Lenders commerce Chain Data Based Banks Giants Financing Lenders Psychometric testing x As noted, however, banks are beginning to take notice of Enabling policy and regulatory developments to support these new players and are increasingly developing partner- the use of alternative data for credit decision-making will ships with new SME lenders. Due to the increasingly digital be necessary. It will also be important to implement the nature of the economy and the ability to analyze new and necessary checks and balances. Addressing these issues will traditional data much faster and more efficiently than in the require: past, digital data-based SME lending will only continue to grow and expand. • A review of existing rules and regulations in place in order to improve the availability of reliable data for the purpose While both banks and new digital SME lenders often tried of enhancing SME financing, including access to bank to carve out individual market niches, many players are in- data transactional information.  creasingly coming up to a similar conclusion, that is, that they are all stronger and can grow more quickly through • Facilitating enhancements to improve credit information forging collaborative partnerships. Under the G20 High infrastructure, including SME credit reporting and access Level Principles of Digital Financial Inclusion, policy mak- to data by alternative lenders. ers looking at expanding SME access to finance can help to raise awareness among small businesses about the ad- • Increasing cooperation of various regulators, not only on vantages of processing payments and transfers digitally and a national level but also on a regional and internation- the features of available digital financial services. In addi- al basis — especially to support innovative financing for tion, SMEs can be encouraged to make informed choices SMEs involved in global value chains as well as oversight by supporting the development of tools allowing potential of alternative lenders operating in several markets.  borrowers to compare similar credit facilities (such as price comparison websites).86 • Involving policymakers in market competition rules.  With access to more digital data and the increasing usage of • Understanding the challenges and balancing act required new alternative data, there is also a whole new range of is- to address consumer protection, data privacy and the im- sues for policymakers and regulators that requires attention plications for increased cyber security measures in light of and analysis.  the use of new alternative data, new players and increas- ingly interconnected partnerships.  38 ALTERNATIVE DATA TRANSFORMING SME FINANCE The ICCR can serve as a useful forum for developing glob- partnerships is the use of test-and-learn approaches such al guidance on how to deal with this exploding supply of as regulatory sandboxes or others.87 These would be similar both data and data users. The G20/GPFI should support the to approaches now being developed to facilitate alternative ICCR’s work in this area. lending models in Australia, Hong Kong SAR, China, Indo- nesia, Singapore, and the U.K. Regulatory sandboxes might Another rapidly expanding option for regulators to facili- carry risks and may be detrimental to a level playing field. tate the experimentation of alternative lending models and G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 39 Endnotes 1. For the purposes of this report, the term SME includes for- 6. Woods, Dan. “How Big Data Flows Will Change Business mal and informal micro, small, and medium firms. The mar- Lending” (September 28, 2015). http://www.forbes.com/ gin of error is plus or minus 9.4 percent for the global devel- sites/danwoods/2015/09/28/how-big-data-flows-will- oping market total, resulting in a range of between 360 and change-business-lending/#df8c7d6297ab 435 million SMEs across global developing markets (simi- lar global data is unfortunately not available for developed 7. Screen scraping refers to the process in which customers economies). Formal SME counts are structured consistent give third parties their online banking user names and pass- with that used by the IFC Enterprise Finance Gap database words so that those third parties can log in on the custom- according to the size of employment: micro enterprises 1-4 ers’ behalf and copy and paste their account information employees; small 5-49 employees, and medium 50-249 into other programs. The practice is controversial because employees. Informal firm counts include enterprises that it poses customer data security concerns and the technol- are not registered with the municipality or tax authority and ogy process can cause spikes in online banking traffic that all the non-employer firms (independent of registration). can lead to outages. For more information, read Peter, Brian Counts exclude public administration and not-for-profit or- “If Banks Fear Screen Scraping, Why Are They Fighting the ganizations as well as enterprises active in the agricultural Alternative?” January 4, 2016. https://www.americanbank- sector. For detailed information, data sources, and estima- er.com/opinion/if-banks-fear-screen-scraping-why-are- tion methodologies for SME counts, credit demand, credit they-fighting-the-alternative gap, and deposit gap market information included in this section, see the “Methodology” tab at: https://smefinance- 8. Gorfine, Daniel and John Schellhase. “From Modes- forum.org/data-sites/ifc-enterprise-finance-gap to To Mombasa, Tech Is Revolutionizing Small Busi- Sources: International Finance Corporation (IFC) Enterprise ness Lending” (August 6, 2014). http://www.forbes.com/ Finance Gap Database (2011); Global Payments Experts llc. sites/realspin/2014/08/06/from-modesto-to-momba- (GPE) analysis. sa-tech-is-revolutionizing-small-business-lending/ 2. Unmet credit needs include SMEs that need credit and are 9. McKinsey & Company (March 2012). “Micro-, small, and not served at all, or that have some credit, but need more. medium enterprises in emerging markets: how banks can Sources: IFC Enterprise Finance Gap Database (2011); McK- grasp a $350 billion opportunity.” insey & Co. (August 2010), “Assessing and Mapping the Gap in Micro, Very Small, Small, and Medium Enterprise (MSME) 10. Global Payments Experts llc., McKinsey & Company, and Finance” and “Two trillion and counting;” GPE analysis. IFC proprietary analyses. 3. Read BBVA Working Paper Impact of capital regulation on 11. IFC (2007). “Benchmarking SME Practices in OECD and SMEs credit (2017) for more on the impact that regulations Emerging Markets.” such as Basel III may have on SME lending https://www.bb- varesearch.com/wp-content/uploads/2017/01/WP-17-01. 12. McKinsey & Company (March 2012). “Micro-, small, and pdf medium enterprises in emerging markets: how banks can grasp a $350 billion opportunity.” 4. “The Digital Universe of Opportunities: Rich Data and the Increasing Value of the Internet of Things” (April 2014). IDC. 13. IFC (2009). “SME Banking Web Survey Report.” 5. “Data infrastructure costs fell by 20 percent per year 2010- 14. “Small Business Banking and the Crisis: Managing 2015.” Source: IDC, May 2016. Development and Risk” (2010). Capgemini, FFMA, and Uni- Credit Group. Based on Capgemini’s analysis from 58 bank 40 ALTERNATIVE DATA TRANSFORMING SME FINANCE interviews globally in 2010. Eighty-one percent of the banks make and verify transactions on a network instantaneous- were in 15 countries in Europe (both Eurozone and non- ly without a central authority. Blockchain is a data structure Eurozone). that makes it possible to create a digital ledger of transac- tions and share it among a distributed network of comput- 15. Quittner, Jeremy. (April 7, 2011). “Nonbanks Leapfrog ers. It uses cryptography to allow each participant on the Banks in Small-Business Cash Management and Treasury.” network to manipulate the ledger in a secure way without Bank Technology News. the need for a central authority. Once a block of data is re- corded on the blockchain ledger, it’s extremely difficult to 16. Kumar, Rajath, “Digital financing: The way forward for change or remove. When someone wants to add to it, par- financial inclusion in Asia” https://blog.capitalfloat.com/ ticipants in the network — all of which have copies of the digital-financing-the-way-forward-for-financial-inclu- existing blockchain — run algorithms to evaluate and verify sion-in-asia-e27/. In addition, the European Central Bank’s the proposed transaction. If a majority of nodes agree that quarterly access to finance surveys since 2009 have consist- the transaction looks valid — that is, identifying information ently shown that a range of two percent to as high as 16% of matches the blockchain’s history — then the new transac- SMEs (depending on country) do not apply for a bank loan, tion will be approved and a new block added to the chain. credit line, or overdraft because they fear a decline. Many financial and other firms across industries see and are experimenting with this distributed ledger technology as a 17. Wehinger, Gert, “SMEs and the credit crunch: Current fi- secure and transparent way to digitally track the ownership nancing difficulties, policy measures and a review of litera- of assets, a move that could speed up transactions and cut ture” OECD Journal: Financial Market Trends Volume 2013/2 costs while lowering the risk of fraud. Some promising appli- https://www.oecd.org/finance/SMEs-Credit-Crunch-Fi- cations include using blockchain to track the movement of nancing-Difficulties.pdf assets throughout supply chains, electronically initiating and enforcing contracts, verifying loan documentation and own- 18. “The Future of Financial Services: How disruptive innova- ership, and much much more. See Norton, Steve (February 2, tions are reshaping the way financial services are structured, 2016). “CIO Explainer: What Is Blockchain?” https://blogs.wsj. provisioned and consumed” (June 2015). World Economic com/cio/2016/02/02/cio-explainer-what-is-blockchain/ Forum (WEF), prepared in collaboration with Deloitte. 25. The marketplace lending category also includes investor 19. BI Intelligence (April 2016). ecosystem functions and firms that provide an array of auxil- iary tools and technology. Together they serve to streamline 20. The Economics of Peer-to-Peer Lending (September the way that investors interact with the P2P lending market- 2016) Oxera Consulting LLP. http://www.oxera.com/get- place asset class. It includes firms (or marketplace lending media/9c0f3f09-80d9-4a82-9e3f-3f3fefe450b2/The-eco- platform capabilities) that target the creation of secondary nomics-of-P2P-lending_30Sep_.pdf.aspx?ext=.pdf markets. This allows investors to sell their loans to other in- vestors and the functions that support it, as well as platform 21. Thomas, Paul, Managing Director. Provenir “Connecting technology and firms that allow lenders to launch compliant the dots: the digital transformation of risk analytics and marketplace lending platforms quickly. Some of these tools, decision-making.” Financial IT, March Issue 2016. technology, reports, and services support institutional inves- tors; others support retail investors. Investors require an ar- 22. Deloitte (May 2016). “A temporary phenomenon? Market- ray of transparent data and loan performance evaluation and place lending.” https://www2.deloitte.com/content/dam/ monitoring, as well as compliance tools to make sound in- Deloitte/uk/Documents/financial-services/deloitte-uk-fs- vestment decisions. The data and tools required to be trans- marketplace-lending.pdf . See also “Deloitte Report Flares parent and compliant with investors are outside the scope Tensions Amid Banks, Alt-Lenders” (May 2016 PAYMENTS) of this report, but are vital to the success of investor-funded http://www.pymnts.com/news/b2b-payments/2016/ marketplace lenders. deloitte-marketplace-lending-report-bank-alternative-mpl/ 26. Cloud-based lending platforms typically provide lenders 23. See the World Economic Forum (October 2015) Report with an end-to-end, modular credit life cycle software plat- “The Future of FinTech: A Paradigm Shift in Small Business form that can be easily configured to meet individual lenders’ Finance”. lending processes; data streams; credit policies; acquisition, scoring, and portfolio management strategies and perfor- 24. Known by many as the technology underpinning the mance objectives; and collections and recoveries strategies. bitcoin digital currency, blockchain is way to let companies G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 41 27. SME loan broker marketplaces are destination websites 36. Renton, Peter (July 21m 2015). “Petralia of Kabbage that attract prospective SME borrowers, and match them Podcast.” http://www.lendacademy.com/podcast-42-kath- to bank and non-bank loan originators that offer their loan ryn-petralia-of-kabbage/ products on the platform. They add value to lenders and borrowers throughout the entire process, often through 37. Interview with Marc Gorlin, co-founder of Kabbage data-driven digital means to provide borrower education (July 19, 2013). https://www.facebook.com/KabbageInc/ and advice about lending options. They also drive volume posts/10151723880621084 to originators, facilitate the lender application process, and (often) utilize sophisticated algorithms that preliminarily 38. See the World Economic Forum, (June 2016). “Innovation credit-qualify and match borrowers to best-fit lenders. The in Electronic Payment Adoption: The case of small retailers”. service is typically free for SME borrowers; instead, the par- ticipating lenders pay fees to the marketplace for borrower 39. Lyon, Ben and Peter Zetterli. (August 3, 2015). “How to referrals. Drive Merchant Payments? Build Solutions Merchants Want.” http://www.cgap.org/blog/how-drive-merchant-pay- 28. See Endnote 7. ments-build-solutions-merchants-want 29. “Interview with Vytautas Zabulis, CEO and Co-Founder 40. The Occupational Safety and Health Administration of Savy” (December 9, 2011). http://www.p2p-banking.com/ (OSHA) is the main federal agency in the U.S. charged with countries/baltic-interview-with-vytautas-zabulis-ceo-and- assuring safe and healthful working conditions, setting and co-founder-of-savy/ enforcing standards, and providing training, outreach, edu- cation and assistance. 30. See Black Economic Empowerment (BEE) in South Africa https://en.wikipedia.org/wiki/Black_Economic_Empower- 41. “Digital Disruption: How Fintech is Forcing Banking to a ment Tipping Point”. (March 2016). Citi. 31. Renton, Peter (January 13, 2017). “Podcast 86: Denise 42. Alibaba Investor Day June 14,2016 company figures; me- Thomas of ApplePie Capital.” http://www.lendacademy. dia and company reports; Global Payments Experts llc. anal- com/podcast-86-denise-thomas-applepie-capital/ ysis. Figures as of March 31, 2016. 32. Psychometrics refers to the measurement of knowl- 43. Xiaoxiao, Li (July 17,2014). “Alibaba Has Big Hopes for edge, abilities, attitudes and personality traits. EFL applies New Big Data Processing Service.” http://english.caixin. psychometric principles to credit scoring by using advanced com/2014-07-17/100705224.html statistical techniques to forecast an applicant’s probability of default. https://www.eflglobal.com/about/faq/ 44. “Alibaba Ups the B2B Ante on SME Access to Working Capital.” (January 24, 2016). http://www.biia.com/alibaba- 33. Yoshimura, Midori (May 28, 2015). “‘It’s A Lot Of Mon- ups-the-b2b-ante-on-sme-access-to-working-capital ey….Enough For Us To Build What We Wanted To Build”: Di- anrong.com CEO On New Valuation of ‘About’ $1 Billion.” 45. Source: Company financials as of March 31, 2016. Global http://www.crowdfundinsider.com/2015/05/68460-its-a- Payments Experts llc. analysis lot-of-money-enough-for-us-to-build-what-we-wanted- to-build-dianrong-com-ceo-on-new-valuation-of-about- 46. As a broader issue, there are various challenges with big 1-billion/ technology giants setting up credit rating services that pol- icymakers and regulators need to be aware of. These tech- 34. FnConn is a subsidiary of the world’s largest contract nology giants have unique data that, from a theoretical point manufacturer FoxConn Technology Group. It provides sup- of view, should be able to yield useful insights into an SME’s ply chain financing to the upstream and downstream suppli- performance and ability to manage both money and busi- ers of Foxconn. FnConn’s specific offerings include financial ness relationships. At the same time, though, if they do not leasing, small loans, business factoring, private equity fund provide the sort of transparency that now exists for credit management and other licenses. scoring providers in developed markets (as in the U.S.), they could cause consumer/SME problems. For example, in the 35. Jackson, Brian (August 29, 2016). “How Kabbage knows if US turndowns could not be explained because the factors it can lend you $100,000 in just seven minutes.” http://www. inside the scoring “black box” were not known. When the itbusiness.ca/news/how-kabbage-knows-if-it-can-lend- scores were opened up, many mistakes in records were you-100000-in-just-seven-minutes/78573 found that, only then, could be redressed. 42 ALTERNATIVE DATA TRANSFORMING SME FINANCE 47. “China Boots Up an Internet Banking Industry”. (January 58. https://bitcoinmagazine.com/articles/how-fluent- 27, 2015). https://www.chinafile.com/reporting-opinion/ wants-to-streamline-financial-supply-chains-with-a- caixin-media/china-boots-internet-banking-industry blockchain-1465318410 48. Paymnts.com (December 2, 2015). “Big Data Uncovers 59. http://wavebl.com Minute Details of Small Biz Borrowers”. http://www.pymnts. com/in-depth/2015/big-data-uncovers-minute-details-of- 60. Lending Startups Look at Borrowers’ Phone Usage to small-biz-borrowers/ Assess Creditworthiness. http://www.wsj.com/articles/ lending-startups-look-at-borrowers-phone-usage-to-as- 49. The information on DHGate’s micro-lending programs sess-creditworthiness-1448933308 and data were obtained through a series of emailed inter- views with DHGate’s CEO Diane Wang and her assistant Ivy 61. Note that the KCB M-Pesa account utilizing the Safar- Zhang in 2015. icom M-Pesa platform is different because it is linked to a bank and hence offers autodebit features. 50. Bose, Nandita (June 29, 2015). “Amazon to offer loans to sellers in China, 7 other countries.” http://www.reuters. 62. A smart loan for people with, as of yet, no credit history. com/article/2015/06/29/us-amazon-com-loans-exclu- http://www.ted.com/talks/shivani_siroya_a_smart_loan_ sive-idUSKCN0P90DW20150629 for_people_with_no_credit_history_yet#t-31346 51. http://www3.weforum.org/docs/IP/2015/FS/GAC15_ 63. Sources: Foundation Capital, Lending Club, McKinsey, The_Future_of_FinTech_Paradigm_Shift_Small_Business_ Liberum. Finance_report_2015.pdf 64. “The Brave 100: The Battle for Supremacy in Small Busi- 52. http://spendmatters.com/tfmatters/a-first-step-to- ness Lending”. (October 2015). QED Investors and Oliver Wy- predictive-analytics-quantifying-the-risk-of-paying-an-in- man. voice-on-receipt/ 65. That said, the early players like Kabbage, OnDeck, Cred- 53. http://spendmatters.com/tfmatters/a-first-step-to- itEase, and Funding Circle, have been consistently lowering predictive-analytics-quantifying-the-risk-of-paying-an-in- their costs of acquisition every year through partnerships voice-on-receipt/ with other fintech players with large SME customer bases like online accounting firms and merchant acquirers as well 54. http://wavebl.com as banks, and through increasingly sophisticated, data-driven targeted marketing programs. 55. Shamah, David (November 3, 2015). “Israeli start-up’s bitcoin-based tech raises a mast for shippers.” http://www. 66. See Endnote 27. timesofisrael.com/israeli-start-ups-bitcoin-based-tech- raises-a-mast-for-shippers/ For an explanation of block- 67. Securitization pools contractual debt (in this case SME chain technology, see Endnote 57. loans) and sells the related cash flows to third party institu- tional investors and banks as securities, which may be de- 56. A blockchain is a data structure that makes it possible scribed as bonds. Investors are repaid from the principal and to create and share a digital ledger of transactions. It uses interest cash flows collected from the underlying debt and cryptography to allow anyone granted access to add to the redistributed through the capital structure of the new financ- ledger in a secure way without the need for a central authori- ing. Securities backed by SME loan receivables are known as ty. Blockchain technologies are therefore well-suited for log- asset-backed securities (ABS). ging and monitoring large amounts of data, such as short- term loans or an industry’s supply chain. Companies like 68. A number of banks are more comfortable in working IBM are now actively exploring how Blockchain can be used with fintechs to solve their immediate pain points, particular- to track transactional data and track this for purposes that ly related to originating a small business loan digitally. These include financing. See http://www.wsj.com/articles/ibm- banks seem less interested in a deeper relationship with a pushes-blockchain-into-the-supply-chain-1468528824 Fintech that might include multiple facets such as a soft- ware platform, marketing support, risk scoring and analyt- 57. https://www.gtnews.com/articles/how-blockchain-can- ics that incorporate alternative data, portfolio management transform-supply-chain-finance/ and servicing, and collections and recoveries. Banks are “go G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 43 slow” and targeted in their focus while many Fintechs mar- 77. There has been a lot of concern over who owns person- ket broader solutions. Unfortunately, banks pursuing an in- al or business data that is now being used to make credit cremental approach may be losing out by failing to leverage decisions. Questions around whether SMEs can access their the full scope of what the Fintechs offer, but they perceive own digital data and grant access to other financial provid- the risk of teaming as outweighing what is to them the still ers, how and what control can an client have over how these to be proven benefits of partnership. See http://ficinc.com/ records are shared as well as the security around how a cli- fintechs-and-banks-an-unequal-partnership-part-one/ and ent’s data is stored. http://ficinc.com/fintechs-and-banks-an-unequal-partner- ship-part-two/ 78. See the Committee on Payments and Market Infra- structures (CPMI) and the World Bank Group (April 2016) 69. http://www.hanwha.com/en/news_and_media/press_ Payment Aspects of Financial Inclusion report http://doc- release/hanwha-group-expands-presence-in-global-fin- uments.worldbank.org/curated/en/806481470154477031/ tech-market.corporate_news.html pdf/107382-WP-REPLLACEMENT-PUBLIC-PAFI-Report-fi- nal-in-A4.pdf 70. Increasing regulatory scrutiny of bank-fintech partner- ships is only likely to increase these due diligence demands. 79. World Bank (March 2017), Credit Information and Firms’ Access to Finance: Evidence from a New Credit-Con- 71. “Neobanks include financial institutions, working, as a strained Status Measure https://www.worldbank.org/en/ rule, only through the Internet without physical offices, as events/2017/03/30/credit-information-and-firms-ac- well as those, specializing in e-commerce. These kinds of cess-to-finance services are dependent on economic and technological de- velopment of the country, adopted legal base and available 80. The massive volume, vast diversity, and real-time nature mature banking. It is in such conditions that arises the desire of emerging alternative digital data types, and the need for to promote and experiment with different banking servic- specialized big data expertise in how each data set is creat- es”, said the deputy director for the department of banking ed and structured, poses such questions as whether a cen- software RS-Bank for R-Style Softlab Maxim Bolyshev. http:// tral database is even advisable or feasible over the long run. eng.banks.eu/news/info/2289/ Accessing such data via API as needed from carefully- vetted member companies or industry consortiums could be 72. See https://clearbanc.com a viable alternate path forward for policy makers to consider. For example, in September 2016, China’s National Internet 73. http://www.techweez.com/2016/05/16/equity- Finance Association (NIFA) launched its Internet Financial banks-equitel-sme-lending/ Industry Information Sharing Platform (IFIISP), which will provide credit data on a loan applicant by sending requests 74. Most industry observers expected much more traction in to all other IFIISP members and collating the results (data bank-digital SME lender partnerships in 2016 and 2017, but around existing loans, amounts borrowed, number of recent few partnerships have been announced so far. At the same inquiries), but without divulging the source data to protect time, cloud-based lending and banking technology firms like competitive insights. The platform is intended to prevent Cloud Lending Solutions, Mirador Financial, Solaris, Mambu, duplicate loan applications, ensure legal compliance, im- Kontomatik, and others have been ramping up their relation- prove information verification processes, and reduce default ships with banks at a fairly brisk pace. rates and business risks for its members. IFIISP members go through a rigorous vetting process before being allowed to 75. See The G20 High Level Principles for Digital Financial join the platform; at launch, the platform included 17 inau- Inclusion (2016) https://www.gpfi.org/sites/default/files/ gural members, including Ant Financial, JD Finance, Lufax, documents/G20%20High%20Level%20Principles%20for%20 and CreditEase’s Yirendai. What’s different about the IFIISP Digital%20Financial%20Inclusion%20-%20Full%20version-. platform is that customer credit data is not centralized, but pdf/ rather inquired upon at the point of loan application via API technology. 76. For additional information on information privacy laws, principles and practices in different regions around the world 81. See Principle #4 Expand The Digital Financial Services see Wikipedia Information Privacy Laws https://en.wikipedia. Ecosystem https://www.gpfi.org/sites/default/files/docu- org/wiki/Information_privacy_law/ ments/G20%20High%20Level%20Principles%20for%20Dig- ital%20Financial%20Inclusion%20-%20Full%20version-.pdf 44 ALTERNATIVE DATA TRANSFORMING SME FINANCE 82. See the Checklist for a Small Firm’s Cybersecurity Program 85. See the WEF report on Innovation in Electronic Payment created by the Financial Industry Regulatory Authority (FIN- Adoption: The case of small retailers and the Committee on RA) to assist small firms in establishing a cybersecurity http:// Payments and Market Infrastructures (CPMI) and the World www.finra.org/industry/small-firm-cybersecurity-checklist Bank Group (April 2016) Payment Aspects of Financial In- clusion (PAFI) report which also examine the role of policy- 83. See the Small Business Borrower’s Bill of Rights See more makers with regard to the legal and regulatory framework to at: http://www.responsiblebusinesslending.org/#sthash. develop a more conducive enabling policy and regulatory sbpSzW35.dpuf/   environment for innovation that supports financial inclusion. 84. In addition, Principle 4 also encourages service providers 86. See Principle #6 Strengthen Digital and Financial Litera- to use multiple sources of digital data for evaluating con- cy Awareness https://www.gpfi.org/sites/default/files/docu- sumer and small and medium enterprise (SME) creditworthi- ments/G20%20High%20Level%20Principles%20for%20Dig- ness. This approach should include appropriate safeguards ital%20Financial%20Inclusion%20-%20Full%20version-.pdf while facilitating development of such data and ensuring a fair, non- discriminatory approach to its use. Examples of 87. For a review of regulatory sandbox approaches being such alternative data sources include mobile phone use, taken see https://www.law.ox.ac.uk/business-law-blog/ utility payments, data enterprise registration information, blog/2016/12/overview-regulatory-sandbox-regimes-aus- and other information that can complement traditional loan tralia-hong-kong-malaysia. It should be noted, however that repayment or insurance-related data. See https://www. regulatory sandboxes need to be developed carefully in or- gpfi.org/sites/default/files/documents/G20%20High%20 der to avoid unintended consequences, see http://www.eu- Level%20Principles%20for%20Digital%20Financial%20In- romoney.com/Article/3645631/Fintech-sandbox-risks-cre- clusion%20-%20Full%20version-.pdf ating-unofficial-endorsements.html G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 45 Annex 1: Digital Lender Business and Alternative Data Profiles P2P SME LENDING PLATFORMS.................................47 Fundation.............................................................................56 ApplePie Capital..................................................................47 GAXFinance and GAX (Growth Accelerator Exchange)............................................................................56 Bitbond.................................................................................47 iwoca..................................................................................... 57 China Rapid Finance (CRF)...............................................48 Kabbage................................................................................ 57 CreditEase............................................................................48 Kopo Kopo...........................................................................58 Dianrong...............................................................................49 Lendingkart..........................................................................59 Faircent.................................................................................50 NeoGrowth Credit.............................................................59 Funding Circle.....................................................................50 OnDeck................................................................................60 Funding Societies............................................................... 51 Square Capital, Square...................................................... 61 Lending Club....................................................................... 51 Thinking Capital.................................................................. 61 LoanZen................................................................................52 Tyro Payments.....................................................................62 Modalku (Parent company is Funding Societies).........52 Waddle..................................................................................62 MoolahSense.......................................................................53 Zoona....................................................................................62 RainFin..................................................................................53 SAVY......................................................................................53 TECH, E-COMMERCE, PAYMENT GIANTS.................. 63 Ant Financial/Alibaba.........................................................63 ONLINE SME BALANCE SHEET LENDERS................... 54 Amazon Lending.................................................................64 AMP Credit Technologies.................................................54 Amazon India......................................................................64 Aprenita................................................................................54 Baidu/Baixin Bank...............................................................64 CAN Capital.........................................................................54 DHgate.com........................................................................65 Capital First..........................................................................55 Paypal Working Capital.....................................................65 Capital Float.........................................................................55 46 ALTERNATIVE DATA TRANSFORMING SME FINANCE Paytm E-Commerce/Paytm Mall....................................65 MOBILE DATA-BASED LENDERS................................. 69 Tencent/WeBank................................................................65 Branch................................................................................... 69 Commercial Bank of Africa (CBA) M-Shwari................69 SUPPLY CHAIN FINANCE PLATFORMS.......................66 Kenya Commercial Bank (KCB) KCB M-Pesa...............70 ApexPeak..............................................................................66 Tala........................................................................................70 Basware................................................................................66 GO Finance..........................................................................66 DIGITAL SME BANKS/LENDERS/ PARTNERSHIPS.............................................................70 Kickfurther...........................................................................66 ABN AMRO...........................................................................70 Kinara Capital......................................................................66 CIBC (Canadian Imperial Bank of Commerce)............70 Remitia..................................................................................67 CivilisedBank.......................................................................70 Tradeshift..............................................................................67 Clearbanc............................................................................. 71 Traxpay..................................................................................67 DBS........................................................................................ 71 Tungsten Corporation/Network......................................67 Fidor/Groupe BPCE........................................................... 71 ANALYTIC/TECHNOLOGY SOLUTION Equitel................................................................................... 71 PROVIDERS................................................................... 68 FINCA.................................................................................... 71 Entrepreneurial Finance Lab (EFL).................................. 68 Heartland Bank................................................................... 72 First Access.......................................................................... 68 Holvi/BBVA........................................................................... 72 Wave...................................................................................... 68 IDFC Bank............................................................................ 72 SME LOAN BROKER MARKETPLACES......................... 69 ING........................................................................................ 72 Business Finance Compared/Bizfitech.......................... 69 JPMorgan Chase................................................................ 73 Funding Options................................................................. 69 Regions Bank....................................................................... 73 Funding Xchange................................................................ 69 Santander............................................................................. 73 Scotiabank........................................................................... 74 State Bank of India (SBI).................................................... 74 Wells Fargo.......................................................................... 74 G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 47 Annex 1 SME Digital Lender Business and Alternative Data Profiles P2P SME LENDING PLATFORMS Bitbond Website: https://www.bitbond.com ApplePie Capital Type of organization: P2P SME LENDER Website: https://www.applepiecapital.com/ Year operations launched: 2013 Type of organization: P2P SME LENDER Headquarters: Berlin, Germany Year launched: 2015 Headquarters: San Francisco, California, United States Active countries/region of operations: Global (because Active countries/region of operations: United States Bitbond leverages bitcoin as a technology and payment network, it is able to operate globally for borrowers and Alternative data utilized: Franchisor data that includes such investors) things as how many stores they have opened or closed over what period of time, what a store should make in volume or Alternative data utilized: Every borrower must connect revenue, what it costs to build, how long it takes to break at least two online accounts in order to complete the even, and other unit economics broken down by geography application. These might be a PayPal, eBay, Amazon, and store footprint size. The franchisor also provides the cri- MercadoLibre, Google Analytics, Debitoor (accounting soft- teria it uses to select franchisees (ApplePie’s loan customers) ware), or a bank account. On Amazon, for example, Bitbond and the market demand size for the product or service in the can see the number of shipped orders and listed items, while franchisee’s store location. a connected eBay account shows the number and quality of feedback received, giving Bitbond a good idea of the Target lending market: Franchisees looking to start, grow, borrower’s business acumen, which correlates with repay- or retrofit their businesses under more than 40 regional ment probability. Bank and PayPal accounts shed light on and national franchisor brands in eight sectors including the financial health of the borrower and show outstanding business and personal services (such as My Salon Suite, V’s loans from other service providers. A Google Analytics ac- Barborshops, and Orange Theory Fitness), quick service res- count shows the traffic and revenue generated by the site. taurants (such as Jimmy John’s and Marco’s Pizza), and con- By connecting such accounts, as well as with social media venience stores (such as 7-Eleven). Franchisees are backed accounts Facebook, Twitter and LinkedIn, Bitbond uses ma- by the proven business models and central resources of the chine learning algorithms on thousands of data points to franchisor brand. assess the creditworthiness of applicants within seconds. In addition, the company uses video verification for both bor- rowers and lenders. Once connected, the user shows the front and back of the passport to the webcam for the securi- ty officer to see, then receives and inputs a personal security code via email to confirm. 48 ALTERNATIVE DATA TRANSFORMING SME FINANCE Target lending market: Global SMEs; the majority are run- shopping frequency. A key part of the company’s strategy is ning some sort of online business, such as a store on eBay, “low and grow” which identifies quality customers and offers Amazon, MercadoLibre, Etsy or their own eCommerce web- them larger, longer-term loans as they demonstrate positive site). To-date, the company has users from 120 countries. credit behavior and allows it to attain significant lifetime cus- The bulk is currently located in the US, Germany, India, or tomer value. the Philippines, but the company has also seen meaningful traction in Brazil, Spain, the United Kingdom, and Canada. China Rapid Finance’s partnership with Tencent, relying on Most investors (both retail and institutional) are currently its popular social chat WeChat app, was able to provide from Germany, Northern America, and the United Kingdom. credit ratings and establish creditworthiness for 50 million Chinese consumers using social networking and comput- er gaming data. CRF’s algorithms examine how long and China Rapid Finance (CRF) how frequently people use Tencent services, from WeChat Website: https://www.crfchina.com to Candy Crush Saga, the popular smartphone game. The Type of organization: P2P CONSUMER LENDER more someone uses social networking services, the more it (SME LENDING IS IN FUTURE PLANS) shows that people are concerned for their reputation, con- Year operations launched: 2001 (P2P lending platform cerned for their integrity. Another key variable turned out launched in 2010) to be internet purchasing history. Even buying points in an Headquarters: Shanghai, China online computer game was another key variable. Another Active countries/region of operations: China unusual CRF finding is that people with drapes on their win- dows, it appears, are more likely to pay their debts. CRF has Alternative data utilized: Analyzes alternatively sourced, on- also introduced pre-approved loans to pre-screened users line social, search, transaction, and browser data. Partners on Tencent’s social platforms. With the pre-selection mod- with China’s various internet platforms, including online el, the lending platform contacts potential borrowers rather travel agencies, online group-buy and shopping platforms, than the borrower seeking out the lending platform. Thus, online gaming companies, online e-commerce platforms, the platforms can actively seek out prospects with desired payment service providers, social networking giant Tencent, characteristics rather than relying on the accuracy of online and online search giant Baidu. As the company analyzes new applications submitted by potential borrowers, which have a data from its data partners and internally produced cred- higher likelihood of fraud. it data from cumulative borrowing behavior on its market- place, it continually refines its credit assessment algorithms Target lending market: Emerging middle-class, mobile ac- and revises data inputs to create a more accurate measure of tive consumers (EMMAs), typically consumers that are often creditworthiness. When borrowers on the marketplace have well-educated and well-employed, but lack credit histories. developed a sufficient lending history with the CRF market- It provides two types of loans: “consumption loans” of be- place, CRF invites them to one of its 107 local data verifica- tween US$72 and US$865 used for consumer purchases and tion centers across China where they can apply for larger “lifestyle loans” which are typically used for larger purchases, loans facilitated on the platform. This simple process in- like education or healthcare funding of between US$865 to volves verifying additional data, including physical data, such US$14,400. The company plans to target SME lending in the as housing, business, and employment. The data gathered future. in connection with the screening and due diligence carried out in the data verification centers is input into the credit as- sessment system, enhancing its credit analytical capabilities, CreditEase including fraud prevention and detection. Website: http://english.creditease.cn/index.html Type of organization: P2P SME AND CONSUMER LENDER; The company employs multiple independent credit scor- WEALTH MANAGEMENT SERVICES FOR HIGH NET WORTH ing algorithms depending on the loan size and terms. The AND MASS AFFLUENT INVESTORS scoring algorithms are highly automated and instantaneous- Year operations launched: 2006 ly produce scoring decisions based on up to thousands of Headquarters: Beijing, China data variables. Some of these variables include a potential Active countries/region of operations: CHINA borrower’s delinquent repayment histories, online behavior, advanced education degree, employment duration with cur- Alternative data utilized: Partner e-commerce platform rent employer, social security status, housing benefits status, seller data, such as transaction volume, shop size, cus- credit bureau records (if any), real property ownership, sta- tomer comments and ratings, and business performance bility of residence, duration of social media usage, and online data on Amazon.cn, Alibaba’s Lazada, Taobao, Tmall, and G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 49 AliExpress platforms, eBay, and Wish; telecommunications, agricultural businesses with one-stop financial services, de- bank account and credit card transaction, insurance, and livering to their financial needs in rural areas. This includes social security data; data from SME customers’ third party services such as credit loan advisory assistance, leasing, software and services providers, such as purchase orders agricultural insurance, rural wealth management, and agri- and invoices in enterprise resource planning systems, logis- cultural machinery financing. tics data, and inventory management data; public data about SME sellers on the internet and social networking sites; sup- ply chain data from upstream and downstream SME sup- Dianrong pliers; interbank and third party payment data; SME budget Website: http://www.dianrong.com planning and accounting data; and SME wealth manage- Type of organization: P2P SME AND CONSUMER LENDER ment data. Year operations launched: 2012 Headquarters: Shanghai, China CreditEase relies on these data sets along with ten plus years Active countries/region of operations: China; Asia of accumulated credit, fraud, and other CreditEase data as- sets to grade borrowers for risk, and then apply risk-based Alternative data utilized: Dianrong uses multiple third- pricing based on the borrower’s risk grade. In a new joint party data, big data analytics modeling, and a highly flexible, venture launched in May 2017 with global business com- configurable risk-control approval system. It utilizes third merce platform Tradeshift, CreditEase delivered a trade fi- party data on transactions, consumption, career, behavior, nancing app that will bring low-cost financing to Chinese travel history, education, and e-commerce data in its credit businesses incorporating electronic invoice data that will scoring models, and applies self-updating machine learn- later expand to upstream and downstream supply chain pro- ing algorithms to the data, obtaining customer permission cesses and data. as needed. For example, it asks borrowers to allow it to purchase data from payment company China UnionPay to CreditEase also established and runs Beijing Zhicheng Credit assess cashflow and can then lend using future income as Service Co., Ltd. (Zhicheng Credit) which specializes in credit collateral. It also uses giant Ant Financial’s credit scoring ser- information and ratings for small businesses (much like Dun vice, Sesame Credit, and links and analyzes information from & Bradstreet in the U.S.), and also launched a risk manage- potential clients’ social media accounts like Weibo, China’s ment cloud platform for the Internet Finance sector in 2016. equivalent to Twitter. In addition, it builds a “knowledge As of May 2017, over 600 companies in the alternate lending graph” of a person’s network and mines the data to ascertain sector covering P2P lenders, marketplace lenders, insurance potentially risky relationships. Apart from analyzing social companies, consumer finance companies, and manufactur- media usage, clients who miss a payment may be poten- ing member companies share their clients’ credit data with tially embarrassed online because Dianrong.com can post each other for free. The shared infrastructure enables faster public requests on the social media site to demand recovery. credit decision making and helps detect simultaneous mul- In addition to Alibaba, Dianrong is also partnering with eBay tiple applications or fraudulent applications. CreditEase has to lend money to Chinese businesses that sell goods to U.S. contributed data for 12 million loans on the platform (the customers on eBay. largest contributor to the database). As of February 2017, the platform had already identified 900,000 borrowers with More recently, Dianrong partnered with FnConn (a sup- loans or application attempts on multiple online lending ply chain financing subsidiary of the world’s largest con- sites, including over 80,000 who applied for a loan on five or tract manufacturer FoxConn Technology Group) to launch more different platforms at the same time. Chained Finance, a new blockchain platform for supply chain finance. The platform records and authenticates every Target lending market: Small businesses and consumers payment and every supply chain transaction, creating great- for lending; high net worth and mass affluent investors for er visibility into suppliers and their trading data for SME lend- wealth management. CreditEase is also the parent compa- ing decisions. ny of online P2P consumer lending platform Yirendai, which targets salaried workers, launched in 2012, and went pub- Dianrong is also integrating blockchain technology across lic on the U.S. stock market in December 2015. CreditEase its entire platform, where appropriate, to further enhance sub-segments its lending business to target urban salary transparency and security for borrowers and lenders. In workers, small and micro entrepreneurs, and college stu- what the company calls ‘D-Chain’, Dianrong’s blockchain dents for advisory services and loans; it also provides car technology is based on Ethereum with smart contract ca- loans and mortgages. In addition, the agricultural and farm- pabilities, allowing a number of applications under private ing unit of CreditEase Inclusive Finance provides farmers and or consortium chain settings. Applications of Dianrong’s 50 ALTERNATIVE DATA TRANSFORMING SME FINANCE blockchain technology so far include electronic contacts or her debt after spending recklessly and hitting his maxi- (contracts stored on blockchain to prevent tampering) and mum credit limit will rank poorly on the site. Faircent also credit management (ensuring the validity and timeliness of evaluates borrowers’ lifestyle and spending patterns (for ex- borrower credit data via recording and storage on block- ample buying the latest phone or frequenting a pub, among chain). Dianrong invites partners from different industries to others) by analyzing the bank account and payment trans- join Dianrong’s blockchain ecosystem, which it is exploring action data it collects, evaluates current repayments and providing free in exchange for getting data in return from monthly obligations of the borrower, and assesses soft data them to help build its own online lending business and cus- like family details, residence and office stability. tomer verification. Faircent’s credit policies are quite conservative: the com- Target lending market: Specializes in consumer and SME pany’s tech-enabled algorithms reject over 90 percent of loans, ranging in size from about US$365 to US$72,500 for the borrowers who apply for loans, as Faircent focuses on personal loans and about US$7,300 to US$290,000 for SME responsible growth. Faircent generally gives alternative data loans. For banks and companies, Dianrong provides “in- about 50 percent of the risk weighting in its credit scoring frastructure-as-a-service” banking solutions in China and models; the rest of the weighting relies on the borrower’s around the world. Based on the same technology that drives CIBIL1 credit data and scores. That said, Faircent does lend to its P2P lending marketplace platform, the solutions are fully companies and consumers that have no CIBIL credit record modularized and customizable for clients, and span lender if there is sufficient predictive alternative data to do so. and borrower acquisition, payments, big data risk control, data warehousing, credit portfolio management, delinquen- Target lending market: SMEs (34% of Faircent’s borrowers) cy and collections management, and blockchain. and consumers. Faircent Funding Circle Website: https://www.faircent.com Website: https://www.fundingcircle.com/ Type of organization: P2P SME AND CONSUMER LENDER Type of organization: P2P SME LENDER Year launched: 2014 Year launched: 2010 Headquarters: Gurgaon, India Headquarters: London, United Kingdom Active countries/region of operations: INDIA Active countries/region of operations: Germany, The Neth- erlands, United Kingdom, United States. Alternative data utilized: Faircent is technically integrated with Transunion, Yodlee, Lenddo and Jocata. Yodlee pro- Alternative data utilized: Funding Circle auto-populates vides the bank scraping technology which lets Faircent, underwriting information through automated program- with SME permission, to digitally link and update multiple ming interfaces (APIs) from many data sources into its credit borrower deposit and credit bank account transactions in decision engine, allowing it to make very fast decisions lieu of providing bank statements; Transunion helps with the with more flexibility than traditional lenders. While it uses Aadhaar-based eKYC and bureau scores; Lenddo helps with traditional business metrics - including business cash flow, social media scoring; and Jocata provides income tax pulls. personal cash flow, collateral, and personal assets that could The company has a rules engine that breaks the borrowers be liquidated if necessary - Funding Circle expands its pop- and buckets them into different profiles. The bucketing is ulation of eligible borrowers by examining alternative data, done based on a scoring model (out of 400). Faircent eval- including real-time cash flow, Yelp reviews, and an owner’s uates each borrower that registers on the basis of his or her passion about the market opportunity. It obtains alternative ability, stability, and intention to repay across more than 55 SME data, insights, and customers through its SME target parameters from the data it collects. With borrower permis- and data-rich referral partners, including Santander and RBS sion, it also collects alternative data from social platforms in the UK (banking and payment transaction data), software such as LinkedIn and Facebook to supplement its analysis, firms Intuit and Sage (cloud accounting and business finan- although it is still evaluating this data to prove how much cial data) and H&R Block (bookkeeping, payroll, taxes and incremental fraud or credit risk screening prediction it offers. other accounting data). It also takes personal credit history The platform’s algorithms also detect good and bad credit and financial stability of the business owner into very serious behavior. For example, a borrower struggling to pay back his consideration, and requires a personal guarantee from each applicant. 1. Formerly Credit Information Bureau India Limited (CIBIL), Transunion acquired an 82 percent stake in CIBIL to become Transunion CIBIL Limited. G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 51 Funding Circle’s examination of social media includes a Funding Societies Google search that analyzes product or service offerings, Website: http://www.fundingsocieties.com/ management experience, trading history, partnerships, Type of organization: P2P SME LENDER transparency, corporate citizenship, diversity of clientele, Year launched: 2015 seasonality, customer experience, and more. Such informa- Headquarters: Singapore tion can lead to a faster decision on loan applications and/ Active countries/region of operations: Singapore, Indone- or a lower interest rate. Key business indicators revealed sia (Moldaku brand), Malaysia (Moldaku), Southeast Asia through social media also include: Alternative data utilized: The company appears to use tra- • Customer reviews and engagement: Strong and active ditional financial statement lending processes at this time. relationships with existing and prospective customers on social media can be reliable, revenue-drivers for business- Target lending market: SMEs es – and a signal to lenders that the company has a loyal customer base. Lending Club • Customer service: A review of Facebook and Twitter activ- Website: https://lendingclub.com ity to see what customers are saying about the SME, and Type of organization: P2P CONSUMER AND SME LENDER how quickly and effectively the SME responds to com- Year launched: 2006 ments  and  complaints. Quick and friendly customer ser- Headquarters: San Francisco, California, United States vice is a good indicator of an SME’s future success with Active countries/region of operations: United States customer retention. Satisfied customers are valuable for word-of-mouth marketing purposes. Alternative data utilized: The company uses proprietary al- gorithms that leverage behavioral data, transactional data, • Thought leadership: Social profiles that are up-to-date; and employment information to supplement traditional risk blog about relevant topics in the SME’s industry; and the assessment tools, such as Fair Isaac Corporation (FICO) level of those posts shared on Facebook, Twitter and scores. It also uses a combination of third-party data, so- LinkedIn that come across as useful information, rather phisticated analytical tools, and current and historical data than a sales pitch can show the SME owner is perceived as obtained during the loan application process to help deter- an expert or pillar in the SME local community. mine fraud risk. The company does not reveal its data sourc- es or alternative data used in its proprietary scoring models. Target lending market: Established SMEs in business at least Most analysts believe they include data sourced from public two years (and profitable in at least one of those two years). utilities, social media, and possibly other financial service firms such as Mint.com, which has access to thousands of As of yearend 2016, Funding Circle topped US$3 billion in banking profiles and maintains a partnership with Lending SME lending since launch with US$1.4 billion going to small Club. business in 2016. Funding Circle also reported it had reached profitability in the UK as year over year growth surpassed 90 More is known, however, about the alternative data Lend- percent. ing Club uses in two of the three major business lending partnerships it has struck with Google, Alibaba, and Sam’s Clubs. Google uses Lending Club to provide low cost two- year loans up to $600,000 to its Google for Work network of more than 10,000 partners, including resellers, consultants, and system integrators which help Google distribute its ap- plications and services and invests in their growth by fund- ing the loans from its own cash. Google already knows the borrowers and provides that data (such as sales, contracts, income, identification information) to Lending Club; Lend- ing Club crunches the data with a customized underwriting model to evaluate the Google SMEs’ creditworthiness and services the loans. 52 ALTERNATIVE DATA TRANSFORMING SME FINANCE In the Alibaba “Alibaba.com e-Credit Line, Powered by Lend- print data.”2 LoanZen uses the invoices SMEs supply to the ing Club” collaboration, Lending Club gained exclusive rights platform to determine the volume and tenor of the expect- to finance U.S. SME purchases from Chinese suppliers on ed cash flows against which to extend unsecured credit. the Alibaba.com. Lending Club gets direct access to Alib- Borrowers also connect their accounting, tax, and online aba’s large base of U.S. buyers and sellers and their online banking data to LoanZen’s artificial intelligence-based sys- transaction and trade data with Chinese suppliers to more tem, which completes the credit assessment within 15 min- accurately assess risk, make decisions faster (in five minutes), utes. LoanZen also uses data supplied by partner companies, at a lower risk, and at lower interest rates than SMEs typically such as Treebo, a technology-and-analytics-enabled hotel can secure. With direct access to Alibaba’s SME trade and chain. LoanZen uses Treebo’s data on their partner hotels, sales data, Lending Club can match repayment terms to the such as past booking history, future bookings for the prop- cash-flow cycles of borrowers, vet suppliers and shipments, erty, guest feedback collected digitally, and quality perfor- and transfer the funds directly to the suppliers. In the Sam’s mance data Club deal, Lending Club became the launch partner of Sam’s Club Business Lending Center, a fast, simple online platform Target lending market: Caters only to private limited com- connecting members with responsible lenders and offering pany borrowers in services and manufacturing that have savings of 20 percent on loan fees (funded by Lending Club large, reputed clients (multinational corporations [MNCs] or in lieu of a referral fee). Lending Club provides members with listed Indian corporates). SMEs across these sectors furnish access to term loans of up to $300,000 with low rates and their invoices on the platform and receive unsecured loans affordable fixed monthly payments. It is not clear what, if any, from accredited investors who are comfortable with the data Sam’s Club shares about its 600,000 SME members. borrower credit profile. Target lending market: Consumers and SMEs (in business for at least 24 months with at least US$75,000 in annual Modalku (Parent company is Funding Societies) sales). The company also provides auto, student, and patient Website: https://modalku.com financing loans. SME loan products include loans and lines Type of organization: P2P SME LENDER of credit that are fixed or variable rates in amounts ranging Year operations launched: 2016 from US$5,000 to US$300,000 for terms of three months to Headquarters: Jakarta, Indonesia five years. Active countries/region of operations: Indonesia As of December 2016, since launching business loans Alternative data utilized: Analyzes potential borrowers by in March of 2014, the company has made more than US going through five steps in the screening process, which $5 billion in SME loans. include a profile screening, an anti-fraud verification with a site visit, and a psychometric credit information tool test facilitated by the financial technology company Entrepre- LoanZen neurial Finance Lab (EFL). Website: https://loanzen.in Type of organization: P2P SME LENDER Target lending market: To be eligible for a loan, businesses Year operations launched: 2015 must have a turnover of at least US$1,500 per month with an Headquarters: Bengaluru, India operational history of at least 2 years. Modalku offers loans Active countries/region of operations: India to SMEs ranging from about US$3,750 to US$37,500 with a tenor of three, six and twelve months claimed to be ready Alternative data utilized: “All the data that we use in decision for disbursement within 10 days. making is collected online or electronically, from borrowers and relevant public and private sources,” said Amit Gupta, who leads data science at Bengaluru-based LoanZen. “Typi- cal sources are banking data, government data on company identifiers, financials, marketplace data, [and] social foot- 2. http://economictimes.indiatimes.com/small-biz/startups/only-clean-social-history-can-get-small-companies-a-loan/articleshow/53533416.cms G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 53 MoolahSense but also additional non-traditional data points. These can in- Website: https://moolahsense.com/ clude procurement history and social media that will be used Type of organization: P2P SME LENDER to assign the SME to one of seven risk-rating levels. Year launched: 2014 Headquarters: Singapore Target lending market: Consumers and SMEs with micro Active countries/region of operations: Singapore; plans and non-registered businesses borrowing under the “per- expansion in Southeast Asia sonal” category, and registered SMEs borrowing under the “business” loan category. Alternative data utilized: The company uses traditional SME screening methods, based on financial statements, bank account statements, company records, and other docu- SAVY mentation. The company does not appear to be using any Website: https://savy.lt/ alternative data at this time. Type of organization: P2P CONSUMER AND SME LENDER Year launched: 2014 Target lending market: SMEs. To qualify for a business Headquarters: Vilnius, Lithuania loan of three to 24 months, the business must have at least Active countries/region of operations: Lithuania for US$217,000 in revenue and trading for two years or at least borrowers; Europe for investors one year of filing with ACRA.3 To qualify for invoice financing of up to 80 percent of the invoice value (minimum invoice Alternative data utilized: SAVY checks borrowers’ credit his- amount is US$14,500) for a term of 15 to 90 days, the busi- tory, reliability, and debt-to-income ratio. But credit scoring ness must have at least US$72,000 in revenue and at least 12 also includes an evaluation of Internet behavior, history, pre- months’ operating history. vious loan requests, and other statistical data (for example, borrower behavior on the SAVY platform which can reveal intentions of the person). If the person does not read in- RainFin formation thoroughly and does everything quickly by using Website: https://www.rainfin.com copy and paste functions, it means he or she does not think Type of organization: P2P CONSUMER AND SME LENDER of repaying the credit and instead the goal is to get money as Year launched: 2012 soon as possible. If a person spends more time on gathering Headquarters: Capetown, South Africa information and modeling different scenarios, he or she ac- Active countries/region of operations: SUB-SAHARAN tually plans repayment of the credit. AFRICA Target lending market: SMEs and consumers in Lithuania. Alternative data utilized: RainFin provides SME loans in an Unsecured and secured consumer loans, real estate mort- innovative partnership with M2North, a company that ena- gage loans, and business loans. bles SMEs and large industrial companies to exchange pro- curement documents. It acts as an electronic intermediary between large companies and their supplier base. SMEs registered with M2North opt in with RainFin to share their existing data, thereby enabling RainFin to assess the credit- worthiness of a business, much like performing credit checks. This provides a risk rating on the individual borrowers using its site. RainFin uses the data to calculate things such as a business’s estimated cash-flow. Also, since many SMEs us- ing M2North have supplied large corporates, it can also ac- cess a firm’s black economic empowerment status and VAT registration. Rain Fin’s online application process includes an intelligent SME-specific credit scorecard that reviews not only an applicant’s transactional history and financial health, 3. Accounting and Corporate Regulatory Authority (ACRA) is Singapore’s national regulator of business entities and public accountants, commonly known as the Registrar of Companies in other countries.  54 ALTERNATIVE DATA TRANSFORMING SME FINANCE ONLINE SME BALANCE SHEET LENDERS Target lending market: Mobile app developers with at least six months of operating revenue for at least one active AMP Credit Technologies app. Aprenita typically lends US$50,000 to US$300,000 at Website: http://amp-creditech.com interest rates of between six and 20 percent over six to 18 Type of organization: ONLINE SME BALANCE SHEET months. Younger companies with at least $5,000 in month- LENDER/CREDIT TECHNOLOGY SOLUTIONS PLATFORM ly revenue can receive weekly advances on earned revenue Year operations launched: 2010 from the App Store and advertising networks at fees rang- Headquarters: Hong Kong, China ing from two to five percent. The advances are repaid when Active countries/region of operations: Hong Kong, the the developer receives revenue from those sources, usually Philippines, Singapore, and the United Kingdom, with within 45 to 60 days. expansions plans across Asia and Europe Alternative data utilized: Combines credit modeling with CAN Capital daily cash flow data enabling established financial institu- Website: www.cancapital.com tions to offer an unsecured SME loan product. AMP’s tech- Type of organization: ONLINE SME BALANCE SHEET nology, which AMP licenses to banks and non-bank lenders LENDER as a software-as-a-service model, allows them to offer and Year operations launched: 1998 manage a new unsecured short-term working capital lend- Headquarters: New York City, New York, United States ing product to these SMEs. Additional information includes Active countries/region of operations: United States analysis of transactional data including electronically veri- fiable cash flows, such as card payments. The technology Alternative data utilized: As a technology-powered finan- studies and measures the cash flows against sector models cial services provider, CAN Capital’s CAN Capital’s Daily Re- of payments and collections to assess credit. mittance Platform™ and its proprietary risk models provide the company with valuable information about the strengths, Target lending market: SMEs for lending and banks for the risks and day-to-day operations of U.S. small businesses. platform. While it initially started as a merchant cash advance com- pany (which is a purchase and sale of a future receivable) with daily repayments deducted from sales, in 2010 it began Aprenita offering term loans as well. It obtains daily merchant sales Website: https://www.aprenita.com/ data from its payment acquirer partners such as World Pay Type of organization: ONLINE SME BALANCE SHEET and iPayment. LENDER Year launched: 2015 Since it has history dating back to 1998, it can also look at Headquarters: New York City, New York, United States its own experiences across certain industry codes and juxta- Active countries/region of operations: United States pose the SME’s cash flows against performance in the indus- try for reasonability. CAN Capital also goes through various Alternative data utilized: Aprenita sources alternative data verification steps to make sure that it can service the loan on their borrowers through direct integration with App properly: is it a real business real? does it have positive cash Stores (GooglePlay and Apple) and analytics accounts (for flow or not? These are determining factors in making the example, Flurry, Localytics, MixPanel, and AppsFigures). Data credit and credit line/loan amount size decision. Merchants Aprenita accesses in the mobile apps market includes cus- can link their accounting systems for CAN Capital to obtain tomer engagement, sales, marketing conversion, outstand- this information or provide financial statements or other ing invoices from the app store, and customer feedback, documentation. In addition, CAN Capital’s 20-year historical among other data. The app stores also provide information experience (covering several recessions) and high lending about the number of times the app has been downloaded, as experience volume gives them a unique perspective to see well as reviews by customers. Advertising networks provide and act on leading indicators of an economic down cycle to data on the amount of revenue the app is generating. Other reduce risk ahead of most other newer digital SME lenders. data analytics platforms reveal how many active users the app has, how much time users typically spend on the app, Target lending market: SMEs, with a concentration in retail, and data on the effectiveness of the app’s marketing efforts. restaurant, grocery, health care, and professional services (in The company does not require personal guarantees or rely other words, businesses that accept electronic payments). on FICO scores to determine a company’s creditworthiness. CAN Capital offers merchant cash advances, direct loans G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 55 with fixed terms and payment amounts, a bank-like install- In February 2016, Amazon India Seller Services launched an ment loan with longer terms, and a flexible payment fea- invite-only seller lending program with Capital First offer- ture loan called Trakloan. The company also partners with ing SMEs on its platform working capital loans ranging from merchant acquiring and other payments companies, such US$7,750 to US$310,000. as WorldPay and iPayment, to offer financing options to the partner’s merchants. Capital Float As of May 2017, average funding amount across the port- Website: http://www.capitalfloat.com folio on its flagship products is about $50,000, and average Type of organization: ONLINE SME BALANCE SHEET term is 14 months; the company has made over US$6.0 LENDER billion in loans to SMEs since launching. Year operations launched: 2013 Headquarters: Bengaluru, India Active countries/region of operations: India Capital First Website: http://www.capitalfirst.com Alternative data utilized: Capital Float targets the popula- Type of organization: FINANCING COMPANY/ tion of SMEs in India that are unable to get loans from banks E-COMMERCE PARTNERSHIP but actually have a significant data footprint because of their Year operations launched: 2012 engagement with the formal economy. This could be data Headquarters: Mumbai, India garnered by selling online or Government data like Aadhaar Active countries/region of operations: India or credit data and scores from CIBIL (Credit Information Bu- reau (India) Limited) or ICRA Limited (formerly Investment Alternative data utilized: For smaller loans, Capital First uses Information and Credit Rating Agency of India Limited). The a tech-driven loan appraisals approach with detailed algo- digital footprint of the SME it may use includes ecommerce rithms to assess a loan applicant’s repayment capabilities platform sales, seller reviews, and rejections; card and mobile and attitudes which take into account many non-financial transactions at point of sale for retailers; and trading data on attributes of prospects to determine credit-worthiness and B2B commerce platforms. It also includes online accounting safety profile. Through a process of trial and error, the com- data and purchase ledgers and more unique data such as pany has evaluated a range of alternative data variables to social media credit scoring and psychometrics (for example, determine which are credit-predictive and which are not. applicants are asked questions to judge things such as their One example: how does a customer, who is in the age group ability to scale a business, their attitude toward credit, and of 25 to 30, who is self-employed but married, behave dif- how they compare to competitors). Capital Float overlaps ferently from a person who is 25-30, self-employed but un- alternative digital data with the SME’s profitability parame- married? The company has found marriage status provides ters, such as current product usage levels, industry margins, incremental value in the credit decision. For online retailers, potential future earnings, risk parameters (such as probabili- it considers e-commerce marketplace data; it also lends ty of default). It also uses conventional credit bureau metrics against card receivables for offline merchants. For larger and bank statements, and tailors its credit-scoring model for loans, it still conducts a customized cash flow analysis using each category of potential borrower it serves. financial statements and obtains references from the SME’s customers, vendors, and suppliers. Target lending market: Capital Float lends between US$390 and US$155,000 to SMEs in India, but most loans range from Target lending market: The company provides SME loans, US$10,800 to US$15,400. The interest rate on these loans loans against property, mortgages, two-wheeler loans, and varies from 15 to 20 percent. It offers a range of products durables loans to Indian SMEs and consumers. The compa- including term loans, merchant cash advances (for retailers ny targets retailers, manufacturers, traders, and doctors and that accept cards or mobile transactions), invoice financ- other professionals which have been in business for at least ing, online seller finance, Pay Later loans (a rolling line of three years. Various SME loan products are available from credit), and taxi finance. It tailors products for specific SME US$4,650 to US$116,300 with loan tenures of six months segments, such as Uber drivers, ecommerce sellers, or B2B to 36 months. About 70 percent of Capital First’s custom- commerce buyers, and struck a partnership with Payworld, a er portfolio is SMEs and entrepreneurs (most formal and payment innovator targeting customers in remote locations, informal micro firms); 23 percent consumers; and the rest to build a custom credit model to reach India’s 12 million large businesses. But on the loan value side, the SME and en- kirana, or local neighborhood, stores. trepreneur portfolio is 85 to 90 percent of the loan balances. 56 ALTERNATIVE DATA TRANSFORMING SME FINANCE Capital Float has over 50 partnerships, including with B2C the banking system working through its partners. Its current e-commerce platforms Snapdeal, Shopclues, Paytm, Flip- bank partners include Regions Bank and the BancAlliance kart, and Amazon; B2B e-commerce sites Alibaba.com, consortium, a network of 200 community banks which refer Tolexo, IndustryBuying, and OfBusiness; SME mobile POS customers who may not be the right fit for traditional bank payments providers MSwipe and Pine Labs; car sharing ser- financing to a website built by Fundation. In the BankAlli- vice Uber (vehicle loans for drivers), and cloud accounting ance partnership, the loan may be funded by either Funda- software provider Intuit, each of which provide it with valua- tion or the partner bank based on various factors, including ble alternative credit scoring data. the business owner’s needs and creditworthiness. Goldman Sachs also signed a US$100 million credit facility for Funda- In December 2016, it partnered with IDFC Bank to provide tion in August 2016. digital lending that will focus on SME borrowers who have no access to bank credit, with limited or no documentation In addition to banks and a portion of loans it continues to and without existing credit history. IDFC Bank will gain ac- hold on its own balance sheet, Fundation partners with busi- cess to Capital Float’s digital network of borrowers, thereby ness-service providers such as Wolters Kluwer N.V. to make enabling it to diversify its portfolio of small ticket loans and loans. It also works with the U.S. Department of Commerce’s grow its customer base. Capital Float, in turn, can leverage Minority Business Development Agency to facilitate lending. IDFC Bank’s balance sheet, product innovation, and custom- ization of banking products for this segment of borrowers. GAXFinance and GAX (Growth Accelerator Exchange) As of March 2017, since launching in 2013, Capital Float had Website: http://gaxfinance.com/ and http://gaxworldwide. originated over US$120 million in loans to over 5,000 SMEs. com/ The company’s credit policies remain stringent, approving Type of organization: ONLINE SME FINANCING AND only 20 percent of applications. The company maintains and SUPPLY CHAIN PLATFORM targets a non-performing asset (NPA) proportion of less than Year launched: Soft launch 2015; full launch delayed until one percent of its total loan amount. 2017 Headquarters: Kuala Lumpur, Malaysia Active countries/region of operations: Malaysia first; then Fundation ASEAN region Website: http://fundation.com/ Type of organization: ONLINE SME LENDER Alternative data utilized: Through GAX partner ecosystems, Year launched: 2011 such as a cloud-based e-claims system for automobile Headquarters: New York City, New York, United States workshops and e-commerce marketplaces for online re- Active countries/region of operations: United States tailers, GAX will provide the platform to thousands of SMEs for their day-to-day operations. The data from these eco- Alternative data utilized: Fundation’s technology uses ex- systems provide unique insights into the businesses to un- tensive data aggregation, proprietary datasets and intelligent derstand their performance and growth prospects. So, if a decisioning techniques to predict credit risk and appropri- business approaches GAX for financing solutions, all GAXFi- ately price loans. The power of its offering is in the tools it nance will ask for is the SME’s ecosystem account. Using the offers around online applications and data-intensive credit data available, GAXFinance will assess the businesses without algorithms to partners. Aggregating third-party data in real paperwork and provide innovative financing solutions based time, doing a lot of automation, using disparate data sources, on their respective needs. The platform will process appli- and combining them to determine what kind of risks Fun- cations in minutes and disburse them within days, offering dation and its partners are taking is a major strength of the fast, accessible financing. GAX will also extend its solutions platform. However, Fundation does not disclose which data to growth and mature stage, medium to large B2C retailers sources or data it uses. with payment gateways as well as SMEs offering software- as-a-service. Target lending market: Fundation currently offers busi- ness loans between US$20,000 and US$500,000. The firm’s Target lending market: The newly launched Growth Ac- loans have terms of one to four years, and they carry an- celerator Exchange (GAX) aims to provide SMEs within the nual percentage rates of between 7.99% and 29.99%. Unlike ASEAN region with financing, payments, and logistics solu- other digital SME lenders, however, Fundation has staked its tions through an all-in-one platform, beginning with op- growth in bank partnerships, and is content to stay entirely in erations in Malaysia. Through its GAXFinance arm, it set an the background as an integrated, credit solutions partner of initial goal of disbursing between 10,000 and 20,000 loans G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 57 to SMEs in Malaysia with a loan size of around US$4,700 to Kabbage US$11,700. For high potential SMEs, GAXFinance will con- Website: http://www.kabbage.com sider a maximum loan amount of around US$23,500. The Type of organization: ONLINE SME BALANCE SHEET interest rate charged on the loans will be between 10 to 16 LENDER per cent, a fraction lower than the 10 to 18 percent charged Year operations launched: 2010 by local financial institutions for the same loan amount in Headquarters: Atlanta, Georgia, United States Malaysia. GAX Finance is currently working with government Active countries/region of operations: Global. Currently agencies and business partners across Southeast Asia to pio- active in the United States, Australia, Mexico, Canada, Spain, neer secure partners and launch the GAX ecosystem. and the United Kingdom Alternative data utilized: Analyzes large numbers of trans- iwoca actions for each loan. It is able to do this by pulling alternative Website: www.iwoca.co.uk data from bank accounts, merchant acquiring processors, Type of organization: ONLINE SME BALANCE SHEET social networks, e-commerce sites, accounting software, LENDER shipping records (for example, from UPS), and dozens of Year launched: 2012 other private and public sources to gauge the risk and cred- Headquarters: London, United Kingdom itworthiness of the business seeking the loan. The compa- Active countries/region of operations: United Kingdom, ny’s underwriting engine pulls information such as business Germany, Poland, and Spain; expansion to more countries revenue, vendor payments, and tax and accounting data to in Europe in the pipeline assign the proper line of credit in real time. Alternative data utilized: Uses online payment or POS mer- Merchants can optionally link their Facebook, Twitter, and chant acceptance accounts (for example, Magento, Skrill, UPS shipping accounts, which may qualify them for fee dis- Shopify, Sage Pay, Paypal, Linnworks), as well as online ac- counts. Executives describe this data as the ‘space between counting services (for example, FreeAgent, Sage), business that data’ to decide whether Kabbage is going to offer the bank statements or APIs to business banking account trans- merchant capital. Facebook business, Yelp, Foursquare, Am- actions, VAT returns (which can be downloaded directly from azon, and eBay offer business reviews, rankings, and other the U.K. government’s Her Majesty’s Revenue and Customs rich data on how businesses actually interact with their cus- (HMRC) web site, and company accounts during the appli- tomers. Kabbage has determined that customers who link to cation process. VAT returns provide sales history; company their social media information are 20 percent less likely to be accounts show business profitability; payment and POS ac- delinquent than those who do not. counts capture sales and identity information; and account- ing records provide a comprehensive view of the business Similarly, linkages with logistic and e-commerce providers is financials. iwoca recently also fully integrated with Xero SME producing relevant data. UPS shipping data can reveal how accounting software on top of Xero’s industry leading open many packages an SME is shipping, how many packages are API. The integration enables SMEs already on the Xero plat- returned, the longevity of the business, and if the quantity of form multiple times a week managing their financials who packages shipped is increasing or decreasing. If the compa- decide that an iwoca credit line (up to US$130,000) is right ny knows that someone has been shipping antique mugs for for them to get an approval instantaneously or within a few at least two years for eBay and Amazon, always ships out via hours, and to see the funding immediately reflected in their two-day UPS air, has more than 500 friends on Facebook, accounts. and is always sending out deals on Twitter, then they are often a better risk — regardless of their credit score. Target lending market: SMEs (including startups). Provides credit lines up to US$130,000 in the U.K., US$55,000 in Target lending market: In the U.S., Kabbage’s core product is Germany and Spain, and US$40,000 in Poland. a revolving business line of credit of US$2,000-US$100,000 with six to 12 month terms. The credit line is dynamic; it Through May 2017, iwoca has provided over US$220 million automatically adjusts to furnish businesses with the right to over 10,000 SMEs. amount of capital they need to grow. The average line of credit is US$25,000, and the average client borrower takes 7-8 loans/year and grows revenue by 72 percent. 58 ALTERNATIVE DATA TRANSFORMING SME FINANCE The company began licensing its platform to banks, lend- merchant funds from a centralized pool of funds. Kopo Kopo ers, and other companies beginning in March 2015, and has therefore essentially guarantees loans because merchants since implemented the turnkey, fully configurable platform cannot simultaneously default and continue accepting mo- at ING in Spain, Scotiabank in Canada and Mexico, Santander bile money payments from their customers (they would in the U.K, Kikka Capital online SME lending platform in Aus- have to turn away customers that want to pay with mobile tralia, and Sage and FleetCards U.S.A. in the United States. money to avoid repaying their loan). Second, the technology The three banks will continue their global rollouts in 2017, automatically deducts a percentage of every single mobile and other banks are in the pipeline. It also has major part- money payment in order to amortize an outstanding loan. nerships with United Parcel Service (UPS), Big Rentz (truck As a result, merchants do not have to remember to pay in- rentals), and MasterCard for distribution and data sharing. stallments over the term of their loan; advances are repaid automatically, little by little, every single day. As of April 2017, since launching, Kabbage has extended nearly US$3 billion to 100,000 SMEs across the U.S. and Target lending market: Kopo Kopo is a technology-focused its customers have connected more than 1.4 million data mobile money aggregator and micro-lender which grew out sources to the Kabbage Platform. of an effort to support SME mobile money payments after partnering with Safaricom to acquire merchants to accept M-PESA at the point of sale (this service was later branded Kopo Kopo Lipa na M-PESA, Swahili for “Pay with M-PESA”). The com- Website: http://www.kopokopo.com pany launched Kopo Kopo Grow merchant cash advance Type of organization: ONLINE SME BALANCE SHEET in Kenya in May 2014 for its mobile money merchants. The LENDER company has launched the product in Kenya, Ghana, Tanza- Year operations launched: 2012 nia, and Uganda with bank and mobile network partners, and Headquarters: Nairobi, Kenya will soon launch in Zimbabwe. Active countries/region of operations: Sub-Saharan Africa In 2015, Kopo Kopo repositioned as a merchant acquiring Alternative data utilized: Kopo Kopo Grow is Kopo Kopo’s technology company, offering its advanced payment and merchant cash advance product for mobile money mer- lending technology capabilities to merchant acquirers and chants. The service enables merchants to take cash advanc- other payments companies in emerging markets. The of- es of up to $29,300, with the borrowers paying a 1 percent fering, its Business Operating System (OS), combines mer- fee in lieu of interest. Merchants who repay faster gradually chant-facing services with merchant acquirer tools and get access to larger advances at lower rates. The product scoring algorithms to offer a comprehensive merchant ser- rewards merchants for increasing their acceptance of elec- vices platform to banks, mobile network operators, and third tronic payments, while also giving them access to the capital parties. The technology underlying Kopo Kopo Grow mer- they need to grow. chant cash advance is a core anchor value-added product for merchants included in the platform. Merchant mobile payment data enables Kopo Kopo to build a credit profile that analyzes over 200 variables to price risk In March 2017, Kopo Kopo announced a partnership with and to extend the unsecured loan. The product crunches MasterCard which will see the two organizations roll out hundreds of data-driven “signals” to predict a merchant’s Masterpass QR across 11 markets in Sub-Saharan Africa, im- future cash flow and propensity to default, and then pacting over 250,000 micro and small businesses over the pre-qualifies that merchant for loan range tailored to the next five years. Masterpass QR lets businesses accept mo- business. The merchant then selects the loan amount they bile and digital payments, while reducing their exposure to want, dedicates a percentage of their daily mobile money the risks and costs of managing cash. Payment is instanta- sales to repaying the loan, and digitally signs the terms and neous and guaranteed, meaning that merchants no longer conditions. The whole process from application to loan dis- need to wait days for transactions to reflect in their accounts. bursement takes just minutes. Kenyan merchants will be among the first to benefit - Kopo Kopo will offer the service in partnership with Diamond Trust Several risk mitigation features make the service distinct. Bank (DTB) as part of its acquiring strategy in the country. First, Kopo Kopo (or the acquirer or payment processor) acts as a “master merchant”, which means it reconciles and settles G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 59 Lendingkart Target lending market: Any small business entrepreneur Website: http://lendingkart.com/ with annual turnover of about US$18,500 to US$155,000 Type of organization: P2P SME LENDER to US$232,000 who needs a working capital loan and has Year launched: 2014 Internet access on his smartphone/computer. It takes 15 Headquarters: Ahmedabad, India minutes to apply, the loan is approved in a few hours, and Active countries/region of operations: India the loan is disbursed within three days. The company pro- vides short-term working capital loans ranging from US$740 Alternative data utilized: The company’s credit risk analyt- to US$14,800 at annualized interest rates of between 16 and ics technology analyzes the SME on the basis of over 2,000 24 percent for six to 12 month durations. The average loan variables and data points, which includes industry type, busi- amount is in the US$8,500-US$9,300 range. ness cash-flows and transactions, income tax return filings of the business, and previous loan and repayment records, Since the company plans to reach the remotest parts of the among others. However, rather than ask the customer to fill country to narrow the gap between SMEs and loan provid- out large forms (it only asks for VAT returns and bank state- ers, it aims to overcome the language barriers by issuing its ments), it scrapes the data needed from public and private application forms in various local and regional languages. sources and via APIs information from SMEs’ stores and oth- The company also just recently announced it plans to make er sites. A few examples include accessing seller sales and the credit risk analytics technology it uses to analyze its bor- ratings on ecommerce sites like  Flipkart, Paytm, Voonik, rowers available to other lenders and banks in 2017, provid- and Craftsvilla; data from SmartShift, a mobile app for car- ing it as a “risk analytics as a service” platform. go owners and transporters, enabling users to find cargo transporters based on the shipment size, weight and other As of December 2016, Lendingkart reported a loan applica- requirements; and data from Unicommerce, a multichannel tion approval rate of 22 to 23 percent and the company had order fulfillment platform. In addition to e-commerce SMEs, made over 5,600 loans since launching. Lendingkart services companies that supply goods to large corporations, online app and game developers, and offline retailers and service providers that have at least some level NeoGrowth Credit of online transactions or profiles. Website: https://www.neogrowth.in Type of organization: ONLINE SME BALANCE SHEET In essence, Lendingkart’s strategy is to target all SME niches LENDER in the country where some secondary verified digital data Year operations launched: 2013 is available from third party resources and extract that data Headquarters: Mumbai, India from the SME’s ecosystem to evaluate the company, rather Active countries/region of operations: INDIA than bothering with document submission and evaluation. By doing so, the company says, it is able to obtain enough Alternative data utilized: Analyzes turnover data from a data to determine a customer’s intent to pay back a loan, the merchant’s point-of-sales terminal or online selling his- quality of his product or service, the financial health of his tory. NeoGrowth uses a proprietary credit scoring system business, and ability to survive with competition. to help determine creditworthiness.  It focuses on current and projected cash flows, not financial statements. It has The company also partners with Lenddo to explore and its own filters to assess credit worthiness. Some methods expand the use of social media to help boost lending and it uses are its unique customer acquisition channels (direct credit evaluation. By requesting access to borrower’s social agents, referral Agents and telesales), new payment datasets media account such as Facebook, Linkedin, Google, or Twit- for credit assessment, non-traditional scoring, and dynamic ter, Lenddo can analyze a variety of factors such as who the repayment and automated collections to identify and serve borrower associated with and the reputation and the nature these potential credit-worthy merchants. NeoGrowth con- of his or her contacts. As part of the assessment, Lenddo stantly uses new data to refine its scorecards, and back tests also scans the content available on applicant’s social me- its credit models. NeoGrowth’s entire product proposition is dia page along with the quality of content uploaded by the based on documented historical flows for the last 12 months borrower. Based on the number of factors available, Lenddo at a minimum, as well as a repayment method where Neo can come up with a social media score and Lendingkart can Growth has a direct and independent access to customer’s decide whether and how to weight that information in the cash flows. It also has tailored scoring models for certain lending decision. 60 ALTERNATIVE DATA TRANSFORMING SME FINANCE industry segments. Credit bureau data continues to be an having trouble as well as make changes to the credit score important part of the decision making process, and helps over time. NeoGrowth understand the customer intent/behavior towards repayment. To do so, the company pays far more Purpose-built for SMEs, the company’s technology and On- attention to the underlying data, rather than to the score Deck Score®, now in its fifth generation with each version itself. more predictive than the last, looks at more than 2,000 data points from over 100 external data sources and 10 million Target lending market: The company focuses on meeting SMEs in its extensive internal historical performance propri- financing needs up to US$155,000 of SMEs that are selling etary database to create an accurate business credit profile. goods and services to consumers online or at retail stores that accept credit and debit cards via POS or electronic data Target lending market: OnDeck makes three-to-36-month capture (EDS) terminals. This includes retailers that sell ap- loans of up to US$500,000 to businesses that are often una- parel, consumer durables and electronic items, footwear ble to qualify for traditional credit from a bank. OnDeck ap- and accessories, handicrafts, gifts, food and grocery items, proves or declines the loan within minutes and often funds and optical goods, as well as to restaurants, beauty salons, the loan the next business day. OnDeck can also underwrite hotels, gymnasiums, and health diagnostic centers. Loans the widest credit spectrum compared to its competitors, can extend from 6 months to 24 months depending on the which also reduces acquisition costs. OnDeck has advanced type of loan. its credit model to a point where it can now underwrite a short duration loan to a relatively new business, a line of credit product to a business with sporadic cash flow needs, OnDeck or a bank-like multi-year loan to a mature business. Website: https://www.ondeck.com Type of organization: ONLINE SME BALANCE SHEET In September 2015, OnDeck launched a new QuickBooks Fi- LENDER nancing Line of Credit product powered by Intuit’s customer Year operations launched: 2007 data and leveraging OnDeck’s technology. Intuit SME clients Headquarters: New York City, New York, United States apply for loan offers with a click of a button. In April of 2016, Active countries/region of operations: United States, it licensed its technology to JPMorganChase, which uses it Canada, Australia to offer online SME loans to its own customers. In February 2017, OnDeck partnered with Canada-based small business Alternative data utilized: After filling out a short online financial services and software firm Wave to launch “Lending form, applicants upload cash flow data from cloud-based by Wave” powered by OnDeck available to Wave custom- accounting software (e.g., Intuit, Xero), bank account data, ers in the United States and Canada. Under the partnership, and/or POS merchant payment transactions, which reveals Wave will leverage OnDeck’s lending platform to streamline indicators such as transaction frequency and volume, sea- and automate the business borrowing experience, allow- sonal sales, expenses, and customer revenue. The compa- ing its customers to access OnDeck loans within the Wave ny then analyzes personal and business credit history from ecosystem. credit reporting service providers, scans public and legal re- cords for past lawsuits or liens, reviews OSHA4 records for OnDeck also has small business loan referral partnerships violations, considers the health of an applicant’s industry and with U.S.-based BBVA Compass, OpusBank, and BMO Har- region, and checks online business reviews from sites like ris Bank, and Australia-based Commonwealth Bank. In ad- Yelp, Angie’s List, and Google Places. dition, the company partners with SME solutions providers Angies List and Australia-based online accounting software In order to take into account different types of SMEs, On- firm MYOB (MYOB is also an investor in OnDeck Australia) Deck applies one of a dozen different statistical models to and U.S. payment processors with CardConnect, TSYS, and the data depending on business age, industry and geogra- Heartland as distribution and data partners. phy. OnDeck receives automated electronic repayments from its borrowers either on a weekly or daily basis, thereby As of May 2017, since launching, OnDeck has deployed more providing it with insights into its borrowers’ cash flows. This than US$6 billion to more than 70,000 customers in 700 feature enables it to quickly offer loan adjustments to those different industries across  the United States, Canada  and Australia. 4. The Occupational Safety and Health Administration (OSHA) is the main federal agency in the U.S. charged with assuring safe and healthful working conditions, setting and enforcing standards, and providing training, outreach, education and assistance. G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 61 Square Capital, Square Thinking Capital Website: https://squareup.com/capital Website: http://www.thinkingcapital.ca/ Type of organization: ONLINE SME BALANCE SHEET Type of organization: ONLINE SME BALANCE SHEET LENDER LENDER Year operations launched: 2014 Square Capital, 2010 Square Year launched: 2006 Headquarters: San Francisco, California, United States Headquarters: Montreal, Canada Active countries/region of operations: United States (Square Active countries/region of operations: Canada and Square Capital). Square only: Japan, Canada, Australia, United Kingdom, and Ireland Alternative data utilized: Thinking Capital asks SMEs to link their merchant processing and bank accounts, and obtains Alternative data utilized: Uses machine learning models, data from other sources to make its non-traditional propri- and identifies and makes offers to growing businesses it etary credit decisions. In October 2016, the firm introduced deems credit-eligible. This is based on the SME’s sales and “Lucy”, a new AI chatbot from Thinking Capital who address- payments growth data, the mix of the SME’s new and return- es customer queries via Facebook Messenger 24/7 and helps ing customers (an indicator of how the company grows), the small businesses qualify for financing. Over time, Lucy will daily number and size of sales tickets, and cash flow, among be able to respond more effectively to complex issues as other information. Square data includes US$53 billion in an- her abilities grow through machine learning. Thinking Cap- nual payments volume and more than 2 billion data points ital is also working to integrate Lucy into SMS on their web collected and updated daily on Square merchants (annual- platform so that customers can easily switch from Facebook ized as of March 2017). Messenger to SMS. The focus on mobile is important as 52 percent of Thinking Capital’s customers currently apply for Target lending market: Square Capital provides invite-only loans through mobile devices. loans to Square merchants pitched as a solution to business- es’ cashflow problems, offering finance to help small firms Target lending market: Canadian SMES that accept deb- expand their inventory or otherwise grow their business- it and credit cards across a variety of industries, including es. Square offers the loans to SMEs running their merchant restaurants, retail, auto repair, and health and beauty. The business on Square. Square’s core product allows anyone company provides merchant cash advance and term loan to accept card payments through a reader that attaches to financing of between CA$5,000 to CA$300,000. Businesses a mobile phone or tablet. Square Capital is a merchant loan must have at least average monthly card sales of CA$7,500 issued by its partner bank that advances cash based on fu- and be in business for at least 6 months (term loan custom- ture sales upfront as a lump sum deposited as soon as the ers must have CA$50,000 in monthly sales). next business day in exchange for an agreed-upon fixed per- centage of their future sales as well as a loan fee. Terms are Think Capital’s strategy hinges on partnering with leading flexible and straightforward, and the loan must be paid off technology, retail and financial institutions such as  CIBC, within 18 months of acceptance. The total cost of the loan is Staples, Moneris, Shopify, The UPS Store  and many more a fixed fee rather than an interest rate and the total amount for both distribution and data partners. Key partnerships in- owed never changes. Square restructured the product from clude 1) payments processer and acquirers Moneris, Chase a merchant cash advance into a loan in March 2016 to en- Payments, Global Payments, and TD; 2) Staples to launch able Square Capital to grow faster by selling more loans to Business Loans Powered by Thinking Capital in 2016 in Can- institutional and bank investors, who are more familiar with ada; 3) and Canadian Imperial Bank of Commerce (CIBC) to traditional loans than merchant cash advances as an asset launch a co-branded “Rapid Financing” SME lending plat- class. form and cross-referral partnership in November 2015. As of March 2017, since launching in May of 2014, Square Capital has loaned US$1.5 billion to Square merchants in the United States while maintaining a loss rate of 4 percent. 62 ALTERNATIVE DATA TRANSFORMING SME FINANCE Tyro Payments Books  Online. Business owners that use cloud accounting Website: https://tyro.com are able to link their online accounts to Waddle, and opt-in Type of organization: ONLINE SME BALANCE SHEET for a two-way data exchange automating every aspect of the LENDER financing. Year operations launched: 2003; Tyro Smart Growth Fund- ing launched in 2016 Target lending market: Provides automated, receivables- Headquarters: Sydney, Australia based invoice financing, and a financing add-on. It is a fully  Active countries/region of operations: Australia online, cloud-based platform enabling SME owners to obtain automatic approvals (automated real-time lending) Alternative data utilized: The financing is based on the and ongoing revolving credit lines. The credit line is only for company’s cash flow, its financial health (through the data SMEs that transact B2B. streaming in from Tyro’s POS), and cloud accounting tools used by businesses and linked to Tyro. Zoona Target lending market: Offers Tyro Smart Growth Funding Website: http://www.zoona.co.za as an SME financing service for its merchants. Tyro Pay- Type of organization: MOBILE MONEY-BASED ONLINE SME ments is Australia’s only independent and fastest growing BALANCE SHEET LENDER EFTPOS provider. It serves 14,000+ customers, processes Year operations launched: 2012 over $8 billion annually in card transactions, and has tailored Headquarters: Lusaka, Zambia best-fit solutions for the retail, health and hospitality sectors. Active countries/region of operations: Sub-Saharan Africa. Active in Zambia, Malawi, South Africa, and Mozambique. The company secured a banking license in 2015, and is tar- geting SME financial services. Its new transaction and deposit Alternative data utilized: Analyzes enterprise usage of mo- account solution uses the cloud to automate everything be- bile money transactions. Due to its relationships with its tween the bank, business system, and accounting software. agents and mobile transaction data, Zoona has the ability to By integrating Tyro’s POS data into its banking solutions, carefully produce credit scores for individuals and manage small business owners can obtain a more efficient way to default risks. manage their money and access a loan. Target lending market: Zoona provides financial services through its mobile money platform, and offers emerging en- Waddle trepreneurs an opportunity to provide money transfers and Website: http://www.waddle.com.au financial services to low-income consumers through its net- Type of organization: ONLINE SME BALANCE SHEET/ work of over 1,500 mobile money agents. Initially launched INVOICE FINANCING LENDER in Zambia,  Zoona  has become the country’s leading pay- Year operations launched: 2015 ments service provider and has expanded to Malawi, South Headquarters: Surry Hills, Australia Africa, and Mozambique. The company’s core product is a Active countries/region of operations: Australia, mobile-based Zoona Account that enables entrepreneurs New Zealand to process money transfers, access working capital financ- ing, and manage their businesses. It targets micro and small Alternative data utilized: Uses the SME’s cloud accounting enterprise (MSE) distributors, agents, and retailers in Zambia as well as bank account information. Revolving credit lines with two core products to help them transact with corporate are based on outstanding invoices held in online account- suppliers. ing packages. Waddle links directly into accounting and banking data, enabling it to provide revolving credit lines to Zoona Growth provides an affordable and accessible work- close cash flow gaps and better support business growth. ing capital financing package for Zoona agents linked to cus- Once the accounting application is linked, Waddle calcu- tomer usage and growth of Zoona Payments. Zoona Agents lates a “borrowing base” (the total amount of eligible collat- may pre-qualify for the product and can access larger facili- eral) based on the business data. Waddle then establishes a ties as their payment volume grows; it is particularly popular fluid line of credit to the business. The more the borrower among rural agents. Zoona also works with its agents to pro- uses Waddle, maintains an excellent repayment history and vide other real-time payments products that increase their demonstrates higher sales transactions, the higher the credit turnover and help them grow. limit. Waddle also integrates with  Xero, MYOB and Quick- G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 63 TECH, E-COMMERCE, PAYMENT GIANTS Alibaba and Ant Financial can also spot vendors who have been too aggressive in certain fields and lagging in others by Ant Financial/Alibaba evaluating a mix of data including their promotional cam- Website: http://www.antgroup.com and paigns and profit margins. Based on the results, Ant can pro- http://www.alibaba.com/ vide suggestions as to how such vendors will need to adjust Type of organization: DIGITAL GIANT E-COMMERCE AND their operations and provide financial support accordingly. FINANCIAL SERVICES Another algorithm allows Ant to pace lending more effec- Year operations launched: 2014 (Ant Financial); 1999 tively by increasing credit lines to accelerate inventory pur- (Alibaba) chases needed for big promotions later in the year. This can Headquarters: Hangzhou, China well exceed the typical lending maximum of US $155,000 for Active countries/region of operations: ASIA (but expanding qualified merchants. globally) Ant can support small loan services 24/7 and “310”: users Alternative data utilized: Developed Zhima Credit (Sesame only need 3 minutes to register, one minute to pass the ver- Credit) rating service, which leverages technology and cus- ification process, and then the money is transferred to the tomer behavior analytics. It uses both online and offline data user’s account (zero minutes).5 A typical day of operations to generate credit scores for consumers and SMEs. In addi- would see Ant’s lending program use more than 100 com- tion to owner characteristics, the company takes into con- puting models to process over 80 billion data entries, such sideration the records on sellers, including the number and as the borrower’s credit rating and customer reviews, to form value of their sales, their cash flow through Alipay, comments conclusions about its willingness and ability to repay loans. posted by their buyers, tax payments and customs declara- tions for users who export, shipping and logistics data, and In January 2016, Alibaba forged more than 25 partnerships even utility bills from sellers’ factories. Even vendors that are with credit rating agencies and financial institutions in Chi- in business for only two or three months can secure a credit na and other parts of the world. These new partnerships are line using this wealth of real-time data. enabling Alibaba to better offer SMEs cross-border trade fi- nance. In addition, it enhances their credit rating scoring tool It covers a wide range of fields, including online shopping for SMEs. The service may also help overseas buyers identify habits (e.g. product categories they shop), how people pay trustworthy trading partners and provide Chinese suppliers their bills , and users’ ability to fulfill his/her contract obliga- access to even more financing options. tions, and rates how they use Ant Financial’s financial prod- ucts and services along with their Alipay and investment bal- Target lending market: SME and consumers using the Ali- ances. It also examines the extent and accuracy of personal baba platform. Alibaba’s domestic and international online information (e.g. home address, length of time of residence, business-to-consumer (B2C) and business-to-business mobile phone numbers) and the online characteristics of in- (B2B) virtual marketplaces have 423 million active buyers, teractions between the user and his or her friends and busi- over 10 million active SME sellers, over 40 million SME sellers ness partners. and buyers combined, and, in 2015, generated $485 billion in sales.6 Through its Cainiao logistics affiliate, Alibaba handles Other data that may be considered includes data collected and tracks logistics for millions of parcel deliveries for prod- from social media, e-mails, texts, audio, videos, photos, and ucts ordered on its commerce platforms. Internet logs. In addition, it reviews court reports on peo- ple who deliberately do not repay debts and or return rental The financial arm of Alibaba, Ant Financial, is focused on cars late and data from its partners such as the taxi service serving SMEs and consumers. Ant Financial has 451 million Didi Kuaidi, rating whether users bothered to settle taxi pay- active Alipay users. Ant’s MYBank, a fully fledged internet-on- ments. The company also works actively with financial in- ly bank, launched in June 2015 and in the first quarter 2016 stitutions, various types of merchants, and public agencies merged with Ant Micro Loans, Ant’s SME lending arm, which such as Shanghai Credit Information Services, social insur- provides loans to Alibaba’s merchant sellers up to $155,000. ance, public security and school records to obtain more data that can effectively reflect creditworthiness. 5. Juan, Guo (December 24, 2015).“Has Ant Financial Entered Its Golden Age?” http://www.tmtpost.com/1499061.html 6. Alibaba Investor Day June 14, 2016 company figures; media and company reports; Global Payments Experts llc. analysis. Figures as of March 31, 2016. RMB-USD exchange rate used:6.4834 64 ALTERNATIVE DATA TRANSFORMING SME FINANCE Ant Check Later provides consumer purchase financing up Amazon India to $4,600 on Alibaba’s Taobao and Tmall marketplaces. Website: http://www.amazon.in Type of organization: DIGITAL GIANT E-COMMERCE/CAP- Ant Financial/Alibaba has cumulatively lent $96 billion to ITAL FLOAT AND CAPITAL FIRST LENDING PARTNERSHIPS three million SMEs since 2010 (in January 2016 alone it is- Year operations launched: Marketplace launched in 2013/ sued $4.6 billion in new SME loans.7) Every loan costs 0.3 Credit Facility Launched in 2016 yuan ($.05), roughly one-thousandth of what a traditional Headquarters: Bangalore, India loan by a bank would cost.8 Non-performing loans recently Active countries/region of operations: India accounted for about 1.67 percent of total loans, below the banking industry’s average of 2.34 percent.9 Through this Alternative data utilized: Utilizes SME account tenure, sell- highly efficient system, Ant has been consistently able to er’s history (which takes into account a minimum tenure, control its bad debt ratio below two percent since launching and minimum sales volume), customer feedback and com- in 2010.10 pliance with Amazon policies and guidelines. Target lending market: SME sellers on Amazon India’s Amazon Lending E-Commerce Platform. Website: no website; the program is invitation only Type of organization: DIGITAL GIANT E-COMMERCE Year operations launched: 2012 Baidu/Baixin Bank Headquarters: Seattle, Washington, United States Website: http://www.baidu.com/; Baixin Bank is pre-launch Active countries/region of operations: Amazon is Global; Type of organization: DIGITAL SEARCH GIANT/BANK Amazon Lending is available in the United States, United CO-OWNED BY BAIDU Kingdom, and Japan only. Year operations launched: 2015 Headquarters: China Alternative data utilized: Amazon extends credit to top per- Active countries/region of operations: China forming SMEs on its site by not only analyzing online sales data but also monitoring online end-user feedback from the Alternative data utilized: Using data from Internet search seller’s customer base. This allows Amazon to go beyond the topics, the bank analyzes a potential client’s online history to transactional data to predict how well a business will do in target specific loan products to meet an SME client’s needs. terms of serving its customers demand as well as meeting their needs in a positive and proactive manner. Target lending market: Initial focus on consumer financing, but also looking at SME finance possibilities. Target lending market: Merchant sellers on Amazon’s e-commerce platform. The program is invitation only, and offer amounts have been as little as $1,000 and as high as $600,000. Rates are attractive, ranging from 6 to 14 percent. Sellers approved for a loan may only use the funding for a single purpose: to purchase more inventory to sell on the Amazon marketplace. 7. “Alibaba Group Holding Ltd Affiliate Ant Financial Has Great Plans” (April 27, 2016). http://www.businessfinancenews.com/28729-alibaba-group-hold- ing-ltd-affiliate-ant-financial-has-great-plans/ 8. Xiaoxiao, Li (July 17,2014). “Alibaba Has Big Hopes for New Big Data Processing Service.” http://english.caixin.com/2014-07-17/100705224.html 9. Wu, Cecilia (March 11, 2015). “Alibaba’s MyBank, how does it go in the past 8 months?” http://www.sapidaily.com/alibabas-mybank-how-does-it-go- in-the-past-8-months/ 10. Juan, Guo (December 24, 2015). “Has Ant Financial Entered Its Golden Age?” http://www.tmtpost.com/1499061.html G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 65 DHgate.com Target lending market: PayPal merchants who have been Website: http://www.dhgate.com using PayPal for at least 3 months and processed at least Type of organization: MAJOR E-COMMERCE/BANK $20,000 in sales are eligible to apply for PayPal Working PARTNERSHIP MODEL Capital. Year operations launched: E-commerce site 2004/Bank Partnerships 2012 As of May 1, 2017, PayPal Working Capital has helped more Headquarters: Beijing, China than 115,000 businesses worldwide  access more than Active countries/region of operations: Global $3 billion in loans and cash advances since the ser- vice launched in 2013. Alternative data utilized: Utilizes e-commerce transaction history, personal data, buyer feedback, logistics data, and inventory data. Data analyzed includes factors such as the Paytm E-Commerce/Paytm Mall average number of orders per month, the total transaction Website: https://paytmmall.com/ amounts per month, and the number of disputes received. It Type of organization: MAJOR E-COMMERCE/CAPITAL also includes the duration of being an active seller, number FLOAT DIGITAL SME LENDING PARTNERSHIP of consecutive transaction days, date of first order, buyers’ Year operations launched: E-commerce platform launch- loyalty to this seller, return rate, dispute rate, loyalty rate, and es in 2010; Capital Float seller loans launched in July 2016; so on. e-commerce business spun off and site rebranded as Paytm Mall in February 2017 After a loan is released, after-loan management data analysis Headquarters: Noida, Uttar Pradesh, India is used to detect abnormal behaviors, monitor the process, Active countries/region of operations: India predict trends, make related comparisons, and alert finan- cial institutions if there are identified risks. DHGate’s system Alternative data utilized: Paytm Mall provides seller e-com- ranks specific performance measures, such as sellers’ re- merce transaction, seller ratings, and shipping/logistics and sponse times to questions, product quality, and product in- payments data to Capital Float; Capital Float combines that formation. It also includes information sellers provide about information with its own seller data and algorithms to build the shipping status of orders, potentially making them more a credit profile for the seller and determine the loan offer reliable predictors of business risk and stability. and terms. Target lending market: Partners with banks to provide data Target lending market: Paytm Mall’s 140,000 sellers offer- and financing options for SMEs engaged in cross-border, ing 68 million products to consumers across categories like B2B e-commerce fashion, electronics, consumer durables, and home furnish- ings, among others (as of February 2017). Paypal Working Capital Website: https://www.paypal.com/webapps/mpp/ merchant-working-capital Tencent/WeBank Type of organization: DIGITAL GIANT GLOBAL PAYMENTS Website: https://www.tencent.com/en-us/index.html and COMPANY http://www.webank.com Year operations launched: 2013 Type of organization: DIGITAL GIANT SOCIAL MEDIA/ Headquarters: San Jose, California, United States BANK CO-OWNED BY TENCENT Active countries/region of operations: Live in Australia, Year operations launched: 2015 United Kingdom, and United States. Headquarters: Shenzhen, China Active countries/region of operations: China Alternative data utilized: Paypal draws on insights into how its merchant customers operate. It uses Paypal sales history Alternative data utilized: WeBank’s big data system has 40 data to power rapid lending decisions. It also analyzes eBay trillion records about retail customers. The data include on- seller ratings, whether there is a history of chargebacks or line activities, virtual assets, personal wealth assets, payment too many chargebacks, how the client handles any PayPal behavior, bank account information, purchase activities, disputes, whether there are any holds on the PayPal account, social network information, public credit information, chat and seasonal sales fluctuations to determine credit scores content, a potential client’s social network, gaming habits, for merchants. and so on. The bank also uses data from its many partner companies to assess SME creditworthiness. Target lending market: Consumers and SMEs. 66 ALTERNATIVE DATA TRANSFORMING SME FINANCE SUPPLY CHAIN FINANCE PLATFORMS draw down on a fixed loan amount at the beginning of every week to purchase stock (airtime, mobile-money float, etc.) ApexPeak and pay back the principal, plus interest, at the end of the Website: https://www.apexpeak.com/index.html week, after they have sold their products. Type of organization: SME INVOICE FINANCING Year operations launched: 2015 Target lending market: GO Finance is a microfinance com- Headquarters: Singapore pany that provides working capital to SME distributors in- Active countries/region of operations: East Asia/Pacific, volved in the value chains of multinational corporations that Africa, Middle East manufacture fast moving consumer goods (FMCGs). Alternative data utilized: Conducting background checks on suppliers and buyers. This allows some new models to Kickfurther better manage the risks of financing fake transactions, as Website: http://www.kickfurther.com well as using a data-driven credit scoring engine to assess Type of organization: P2P SME LENDER/ONLINE SUPPLY success or failure rates. CHAIN AND TRADE FINANCE Year operations launched: 2015 Target lending market: Supply chain SMEs. Headquarters: Boulder, Colorado, United States Active countries/region of operations: United States Basware Alternative data utilized: Utilizes transactional data to ana- Website: http://www.basware.com lyze average margin, and annual revenue versus financing Type of organization: ONLINE SUPPLY CHAIN AND TRADE amount requested. Also uses Alexa global rank on traffic FINANCE flowing to the business websites, 3rd party reviews of the Year operations launched: 1985 business and/or their products and services, percentage of Headquarters: Espoo, Finland financing amount covered by existing purchase orders, and Active countries/region of operations: Global the social network outreach of the business on Facebook, LinkedIn, Twitter, and Instagram. Alternative data utilized: Rapid access to timely data about supply chain transactions. Intelligently linked supply chains Target lending market: Inventory financing for SMEs. enable both lenders and suppliers to have the ability to more closely track transactions themselves. They do so by utilizing online dashboards and tools that enable discount financing Kinara Capital arrangements. Website: Type of organization: ONLINE SUPPLY CHAIN AND TRADE Target lending market: Working capital and supply chain FINANCE financing for SMEs. Year launched: 2011 Headquarters: Bengaluru, India Active countries/region of operations: India GO Finance Website: http://www.gofinanceco.com/ Alternative data utilized: Kinara Capital works with net- Type of organization: ONLINE SUPPLY CHAIN AND TRADE work partners (buyers, trade organizations, franchisors, etc.) FINANCE across supply chains to source and fund working capital Year launched: 2011 needs of small manufacturing businesses, artisan clusters Headquarters: Dar es Salaam, Tanzania and agri-retailers. Kinara’s custom risk assessment includes Active countries/region of operations: Tanzania psychometric testing and cashflow analysis as well as flex- ible product terms, customer-centric processes and supply Alternative data utilized: GO provides access to information chain integration that enables the company to complete the technology, or business intelligence, and targeted training loan process in 7-10 days. that assists entrepreneurs to strengthen the capacity of their businesses. It tracks historical sales data from the top-lev- Kinara comes into an existing supply chain and gains an el player in a fast-moving consumer goods (FMCG) value understanding of the money movement and the history of chain (like Airtel or Coca-Cola). GO Finance provides flexible the buyer-supplier relationship: who has been providing to working capital facilities to these clients, allowing them to whom, how much is ordered, how long they have been in G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 67 business, what is the quality of their product, what is the pro- well as their creditworthiness. Also, Tradeshift uses data on jected future demand from this buyer. It uses this informa- the history of payments, how much is due, what it means to tion to make a good credit decision on an unsecured term or cash flow, and other factors. working capital loan. Target lending market: Supply chain buyers and sellers. By focusing on supply chain network partners (such as re- tail chain Mother Earth for artisan, cooperative, and fair trade suppliers, or Villgro for borrowers who are working to reju- Traxpay venate organic farming by buying organic fertilizers, seeds Website: http://traxpay.com and other products in their respective localities), Kinara sub- Type of organization: ONLINE SUPPLY CHAIN AND TRADE stantially lowers the cost of acquisition and credit screening. FINANCE It can then offer loans in 24-26 percent interest range versus Year operations launched: 2009 80-100 percent market rates. Borrower growth allows them Headquarters: Frankfurt, Germany to cover the cost of financing and increase sales and profits. Active countries/region of operations: Global Target lending market: Loans in the range of US$2,000- Alternative data utilized: Online business-to-business US$20,000 to SMEs in India with turnover of less than about supply chain payments from accounts receivable (sending US$310,000, filling the gap between microfinance and com- invoices) to accounts payable (receiving invoices) and mercial capital. cloud-based technologies allow access to improved financ- ing solutions. These avoid many of the challenges associated with more static-based payment models offered by banks. Remitia Website: http://www.remitia.com With cloud-based solutions that offer a 360-degree view of Type of organization: ONLINE SUPPLY CHAIN AND TRADE the business payments along the supply chain, the compa- FINANCE ny has the ability to achieve a flexibility that allows variables Year operations launched: 2015 in the transaction to change. This facilitates supply chain Headquarters: London, United Kingdom financing in real-time and directly connects into and mon- Active countries/region of operations: United Kingdom itors the transaction via an adaptive rules-based engine, which also reduces the risk of lenders that can make use of Alternative data utilized: Transaction data quantifying the this data instantly. risk of paying an invoice on receipt before the usual approv- al process. Statistical invoice modeling allows for a deeper Target lending market: Online business-to-business supply analysis into accounts payable data. This helps to predict the chain SMEs. type of approval risk an invoice submitted from an existing or new supplier may bring. Tungsten Corporation/Network Target lending market: Supply chain SMEs. Website: http://www.Tungsten-Network.com Type of organization: ONLINE SUPPLY CHAIN AND TRADE FINANCE Tradeshift Year operations launched: 2013 Website: http://tradeshift.com Headquarters: London, United Kingdom Type of organization: ONLINE SUPPLY CHAIN AND TRADE Active countries/region of operations: Global FINANCE Year operations launched: 2010 Alternative data utilized: Allows SME suppliers and their Headquarters: San Francisco, California, United States business customers to exchange electronic invoices for Active countries/region of operations: Global financing — without the need for supporting paper trails. Alternative data utilized: Utilizes supply chain transactional Target lending market: Financing for SMEs using an data as well as tracking current cash flow statistics. Tradeshift e-invoice platform. takes into account billing and receivables to assess how long a supplier will take to be paid. This data helps to determine which suppliers to fund, and what rate to offer based on what others similar to them have accepted in the past, as 68 ALTERNATIVE DATA TRANSFORMING SME FINANCE ANALYTIC/TECHNOLOGY SOLUTION Alternative data utilized: The First Access Enterprise Scoring PROVIDERS platform captures data from loan applications, core-banking software, credit bureaus, smartphones and feature phones, Entrepreneurial Finance Lab (EFL) unique commercial partnerships with mobile network oper- Website: https://www.eflglobal.com ators, mobile money platforms, data aggregators, solar com- Type of organization: ANALYTIC AND LENDING PLATFORM panies and other digital product and service providers. The PROVIDER platform turns big data into smart data, separating the sig- Year operations launched: 2010 nal from the noise to delivering simple, reliable information Headquarters: Lima, Peru that financial institutions can use in real-time. The compa- Active countries/region of operations: Global. Currently ny’s scoring technology looks at four kinds of data: demo- active in Africa, Asia, and Latin America.  graphic (age, gender), geographic (urban vs. rural, mobility), financial (airtime, mobile money), and social (call patterns, Alternative data utilized: The score is based on the appli- SMS patterns, number of people). The first two help verify cant’s answers to questions capturing information that can identity and provide context. Financial data provides a predict loan repayment behavior, including the applicants’ pattern of usage, not only mobile money, but also regular attitudes, beliefs, integrity, and performance. EFL analyzes top-up transactions that offer a window into consumption the data to produce a credit score that assesses the appli- patterns. cant’s ability and willingness to repay a loan in real time. EFL uses alternative data such as psychometrics, digital First Access has also respects telecommunications compa- footprints, and cell phone usage information to assess the nies’ desire to safeguard their data by operating on a server repayment risk profile associated with any individual. EFL within their data center, such that no identifiable information continues to improve its psychometric credit scoring capa- leaves their facility. Users must opt into the system before bilities while simultaneously innovating with new alterna- their history can be reviewed in order to facilitate a decision. tive data sources, including mobile phone usage data (via CDR), social network data (via Facebook and Twitter, for Target lending market: Customizable credit scoring plat- instance), and location data (via GPS and GIS). In addition form for lending institutions in emerging markets to credit to its existing online and tablet platforms, EFL has signifi- score anyone  cantly increased its reach by expanding to mobile and SMS. EFL’s credit scoring algorithms are built on top of its grow- ing outcome-based psychometric database. The program is Wave self-learning and constantly incorporates new data to max- Website: http://wavebl.com/ imize predictive power. Type of organization: SUPPLY CHAIN BLOCKCHAIN TECHNOLOGY Target lending market: Lenders serving SME entrepreneurs Year launched: 2014 and the self-employed, and banks and retailers providing Headquarters: Kfar Saba, Israel financing to consumers at point-of-sale. Active countries/region of operations: Global As of May 2017, the EFL database has processed almost one Alternative data utilized: WAVE connects all members of the million applications in 20 languages. supply chain to a decentralized network and allows them a direct exchange of documents. WAVE’s application manag- es ownership of documents on the blockchain eliminating First Access disputes, forgeries and unnecessary risks. Wave has created Website: https://www.firstaccessmarket.com/ a peer-to-peer and completely decentralized network that Type of organization: ANALYTIC AND LENDING PLATFORM connects all carriers, banks, forwarders, traders and other PROVIDER parties of the international trading supply chain. Using de- Year operations launched: 2011 centralized technologies, all communication between these Headquarters: New York City, New York, United States parties will be direct and will not pass through a specific cen- Active countries/region of operations: Global developing tral entity. Due to its decentralized nature, the Wave network markets. Currently active in Africa, Asia, and Latin America. will not have any single point of failure and will not rely on any single entity. Target lending market: Technology supplier to banks and supply chain companies G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 69 SME LOAN BROKER MARKETPLACES to compare rates and options quickly. SME business own- ers answer a few questions and with just one request can Business Finance Compared/Bizfitech reach many funders to get a transparent view of the available Website: https://www.businessfinancecompared.com/ options. Funders respond with individual quotes, bidding for Type of organization: SME LOAN BROKER MARKETPLACE the opportunity to lend and making it easy for businesses to Year launched: 2015 compare loan quotes from over 200 SME lenders. Headquarters: Nottingham, United Kingdom Active countries/region of operations: United Kingdom MOBILE DATA-BASED LENDERS Target lending market: Business Finance Compared, owned by Bizfitech, helps UK SMEs find and compare alternative Branch sources of funding to grow and support their business using Website: https://branch.co innovative technology and analytics. It connects small busi- Type of organization: Mobile Lender nesses with the finance they need to grow in a clear, trans- Year operations launched: 2015 parent and independent way. It tells SMEs what their options Headquarters: San Francisco, California, United States are and what the costs are and importantly, ensures that the Active countries/region of operations: Kenya options presented will result in an approval. The company asks simple questions to match small businesses to the most Alternative data utilized: Credit-scoring models that ana- relevant product for them. It provides an easy way to com- lyze SMS logs, social network data, call data, GPS data and pare finance options, ranks on the facts without bias, and contact lists. searches for the best deals for customers. The platform is not only focused on alternative sources of funding but also Target lending market: Mostly consumers, but also micro offers UK businesses a wealth of independent, impartial in- businesses which actively use their mobile phones to sup- formation and news related content for business insurance, port their businesses. business banking and other business services. Businessfinancecompared.com supported over 50,000 Commercial Bank of Africa (CBA) M-Shwari SMEs in their search for finance and had 40 leading SME Website: http://cbagroup.com/m-shwari/ finance providers on its platform as of November 2016. Type of organization: DIGITAL BANK/PARTNERSHIP MOBILE LENDER Year operations launched: 1962, M-Shwari launched in 2014 Funding Options in partnership with Safaricom Website: https://www.fundingoptions.com/ Headquarters: Nairobi, Kenya Type of organization: SME LOAN BROKER MARKETPLACE Active countries/region of operations: East Africa Year launched: 2011 Headquarters: Liverpool, United Kingdom Alternative data utilized: Customer use of Safaricom mo- Active countries/region of operations: United Kingdom bile and mobile e-money services transaction history. For first-time borrowers, the credit-scoring algorithm consists Target lending market: Funding Options uses sophisticated of a set of Safaricom’s data related to airtime, airtime credit, matchmaking technology alongside expert support to help usage of Safaricom’s e-money product (M-PESA), and the SMEs to find the right finance. length of time the customer has been with the carrier. Each variable has differing weights and scores based on its pre- dictive power. Funding Xchange Website: http://fundingxchange.co.uk The telecom use history of potential new M-Shwari borrow- Type of organization: SME LOAN BROKER MARKETPLACE ers is assessed against these scorecard variables and a score Year launched: 2015 is assigned. The cumulative score of all the variables enables Headquarters: London, United Kingdom CBA to make an informed choice about which new clients to Active countries/region of operations: United Kingdom provide an initial loan to and which to pass on. Repeat loans are then also based on past repayment history. Target lending market: Funding Xchange is a marketplace platform that enables business owners to receive quotes Target lending market: M-PESA mobile e-money from a series of funders from one enquiry, allowing users subscribers. 70 ALTERNATIVE DATA TRANSFORMING SME FINANCE Kenya Commercial Bank (KCB) KCB M-Pesa DIGITAL SME BANKS/LENDERS/PARTNERSHIPS Website: https://ke.kcbbankgroup.com/home/loans/ mobile/kcb-m-pesa ABN AMRO Type of organization: DIGITAL BANK/PARTNERSHIP Website: https://www.abnamro.com/ MOBILE LENDER Type of organization: ESTABLISHED BANK/WORKING Year operations launched: Bank founded in 1896/ CAPITAL CASH MANGEMENT PARTNERSHIP KCB-MPESA loan launched in 2015 Year launched: 1991; in current form, 2009 Headquarters: Nairobi, Kenya Headquarters: Amsterdam, The Netherlands Active countries/region of operations: Kenya Active countries/region of operations: The Netherlands; Global. Alternative data utilized: Utilizes mobile airtime, data top- ups, mobile money transactions, mobile wallet balance, Alternative data utilized: The bank and InvoiceSharing col- age of the applicant, and previous loan status. The use of laborated to launch a comprehensive solution that provides alternative digital data is particularly important in evaluating SMEs with 24/7 insight into their accounts to estimate their first-time borrowers, while repayment-based credit history working capital needs well in advance. An accounting robot becomes more important for subsequent loan applications. tool reads and checks the invoices, generates journal entries, See: “The Proliferation of Digital Credit Deployments” https:// and exports the invoices to the entrepreneurs’ accounts and www.cgap.org/sites/default/files/Brief-Proliferation-of-Dig- accounting system. The robot also compares invoices with ital-Credit-Deployments-Mar-2016_1.pdf historic data from industry partners, using accountancy data based on the preceding three years. SME clients save time Target lending market: M-PESA mobile e-money and money, while ABM Amro builds SME loan volume. subscribers. Target lending market: Small and medium enterprises Tala Website: http://tala.co CIBC (Canadian Imperial Bank of Commerce) Type of organization: MOBILE LENDER Website: https://www.cibc.com/ Year operations launched: 2011 Type of organization: ESTABLISHED BANK/DIGITAL SME Headquarters: Santa Monica, California, United States LENDING PARTNERSHIP Active countries/region of operations: Kenya and Year launched: 1867 (predecessors); 1961 as CIBC Philippines. Plans for West Africa, Latin America, and South Headquarters: Toronto East Asia. Active countries/region of operations: Canada Alternative data utilized: Analyzes SMS logs, social network Alternative data utilized: Merchant processing debit and data, call data, GPS data and contact lists. credit card data, bank account transaction data, and other proprietary data sources used by partner Thinking Capital to Target lending market: Consumers and micro businesses make its non-traditional proprietary credit decisions who are active mobile phone users. Target lending market: SMEs that accept debit and credit cards and have 6 months of operating history CivilisedBank Website: http://www.beingcivilised.co.uk Type of organization: CHALLENGER DIGITAL BANK Year operations launched: 2017 Headquarters: Reading, United Kingdom Active countries/region of operations: United Kingdom Alternative data utilized: Unified back-end database providing a complete profile of all banking; transactional information. Target lending market: SMEs. G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 71 Clearbanc Fidor/Groupe BPCE Website: https://clearbanc.com Website: https://www.fidor.de/ Type of organization: CHALLENGER DIGITAL BANK Type of organization: CHALLENGER DIGITAL BANK/ESTAB- Year operations launched: 2015 LISHED BANK Headquarters: San Francisco, California, United States Year operations launched: 2009, acquired by Groupe BPCE Active countries/region of operations: United States in 2016 (parent company of Banque Populaire and Caisse d’Eparne cooperative banking networks in France) Alternative data utilized: Bank access to online Uber/Lyft Headquarters: Munich, Germany accounts to verify the hours that drivers work, as well as Active countries/region of operations: Global. Active in their average earnings. This allows Uber drivers who are paid Germany, United Kingdom, United States, and Dubai. weekly to receive their daily earnings in advance through a Visa debit card. Use of online accounting systems such as Alternative data utilized: Open API digital banking technol- Intuit’s Self-Employed Solutions accounting software and ogy platform and SME digital banking platform is a key step QuickBooks software for freelancers. toward accelerating Group BPCE digital transformation, and the acquisition will enable Fidor to have a stronger interna- Target lending market: Financial services for self-employed, tional expansion, more technology innovation, and a bigger freelancers, independent contractors, and entrepreneurs. presence in Europe. Target lending market: Global SMEs and consumers. DBS Website: http://www.dbs.com Type of organization: ESTABLISHED BANK/DIGITAL BANK/ Equitel SME LENDING PARTNERSHIPS Website: http://equitel.com Year operations launched: 1968, partnerships with AMP and Type of organization: DIGITAL BANK/MOBILE LENDER P2P lenders, DBS Digibank launched in 2016 Year operations launched: 2015 Headquarters: Singapore Headquarters: Nairobi, Kenya Active countries/region of operations: Asia Active countries/region of operations: Kenya Partnerships Alternative data utilized: Mobile data, calls and mobile • With AMP Credit Technologies (AMP): 2016 launch of e-money transaction data. digital “DBS mLoan” for SMEs • With Funding Societies: 2016 cross referrals with DBS Target lending market: Active mobile phone subscribers. • With MoolahSense: 2016 cross-referrals with DBS Internal FINCA • DBS Digibank: 2016 launch of mobile-only bank with open Website: http://fmh.finca.org/ APIs for consumers and SMEs in India; 2017 launch in Type of organization: GLOBAL MICROFINANCE NET- Indonesia WORK/FIRST ACCESS ALTERNATIVE CREDIT SCORING PARTNERSHIP Alternative data utilized: Transaction data, and electronical- Year operations launched: 1984; First Access/FINCA ly verifiable cash flows to assess the health of the businesses partnership launched in 2016 and its ability to repay (for AMP DBS mLoan only). Headquarters: Washington D.C., United States Active countries/region of operations: Global emerging Target lending market: New approaches focused on ex- markets. Partnership is active in Africa. panding working capital loans to SMEs. Alternative data utilized: Partnership creates a sophisticat- ed alternative credit scoring approach to improve FINCA’s outreach to excluded populations and apply risk-adjusted pricing across the whole spectrum of FINCA’s borrowers in 72 ALTERNATIVE DATA TRANSFORMING SME FINANCE Africa. The First Access Enterprise Scoring platform captures IDFC Bank data from loan applications, core-banking software, credit Website: http://www.idfcbank.com/ bureaus, smartphones and feature phones, unique com- Type of organization: ESTABLISHED BANK/CAPITAL FLOAT mercial partnerships with mobile network operators, mobile SME DIGITAL LENDING PARTNERSHIP money platforms, data aggregators, solar companies and Year operations launched: 2015; Capital Float partnership other digital product and service providers. The company’s launched in 2016 scoring technology looks at four kinds of data: demograph- Headquarters: Mumbai, India ic, geographic, financial, and social. The first two help ver- Active countries/region of operations: India ify identity and provide context. Financial data provides a pattern of usage, not only mobile money, but also regular Alternative data utilized: The digital footprint of the SME top-up transactions that offer a window into consumption Capital Float may use includes data like Aadhaar, ecom- patterns. First Access analyzes existing client data from FIN- merce platform sales, seller reviews, and rejections; card and CA’s operations, as well as subscriber data from local mo- mobile transactions at point of sale for retailers; and trading bile network operators (MNOs) to establish credit scores for data on B2B commerce platforms. It also includes online ac- clients to secure small loans to build businesses or support counting data and purchase ledgers and more unique data emerging personal needs. As First Access collects data, it such as social media credit scoring and psychometrics (for also continuously recalibrates its dynamic FINCA algorithms example, applicants are asked questions to judge things such using machine learning techniques and hands-on collabora- as their ability to scale a business, their attitude toward credit, tion with its data scientists, enabling FINCA to refine its own and how they compare to competitors). product offerings and improve credit quality. Target lending market: Focus will be on SME borrowers who Target lending market: FINCA’s SME and consumer borrow- have no access to bank credit, with limited or no documen- ers in Africa. tation and without existing credit history. Heartland Bank ING Website: https://www.heartland.co.nz Website: https://www.ing.com/en.htm Type of organization: CHALLENGER DIGITAL BANK/DIGITAL Type of organization: ESTABLISHED BANK/KABBAGE SME SME LENDING LAUNCH DIGITAL LENDING PARTNERSHIP Year operations launched: Bank license issued in 2013 Year operations launched: 1991 through merger; Kabbage Headquarters: Auckland, New Zealand partnership launched in 2016 Active countries/region of operations: New Zealand Headquarters: Amsterdam, The Netherlands. Active countries/region of operations: Global. Alternative data utilized: Analyzes bank transactional data, Partnership is active in Spain only. and data from the online accounting software of SMEs. Alternative data utilized: Kabbage pulls alternative data from Target lending market: SMEs and rural lending. bank accounts, merchant acquiring processors, social net- works, e-commerce sites, accounting software, shipping re- cords (for example, from UPS), and dozens of other private Holvi/BBVA and public sources to gauge the risk and creditworthiness of Website: https://about.holvi.com the business seeking the loan. The company’s underwriting Type of organization: DIGITAL BANK/ESTABLISHED BANK engine pulls information such as  business revenue, vendor ACQUISITION payments, and tax and accounting data to assign the proper Year operations launched: 2011, Acquired by BBVA in 2016 line of credit in real time. Headquarters: Helsinki, Finland Active countries/region of operations: Western Europe Merchants can optionally link their Facebook, Twitter, and UPS shipping accounts, which may qualify them for fee dis- Alternative data utilized: Transactions, online accounting, counts. Facebook business, Yelp, Foursquare, Amazon, and and invoice data. eBay offer business reviews, rankings, and other rich data on how businesses actually interact with their customers. Target lending market: SMEs active online. Similarly, linkages with logistic and e-commerce providers is producing relevant data. Shipping data can reveal how many G20 GLOBAL PARTNERSHIP FOR FINANCIAL INCLUSION 73 packages an SME is shipping, how many packages are re- and combining them to determine what kind of risks Fun- turned, the longevity of the business, and if the quantity of dation and its partners are taking is a major strength of the packages shipped is increasing or decreasing. platform. However, Fundation does not disclose which data sources or data it uses. Target lending market: SMEs. Target lending market: Regions SME customers and prospects. JPMorgan Chase Website: https://www.jpmorganchase.com/ Type of organization: ESTABLISHED BANK/ONDECK SME Santander DIGITAL LENDING PARTNERSHIP Website: http://www.santander.com/ Year operations launched: 2000, OnDeck partnership Type of organization: ESTABLISHED BANK/KABBAGE SME launched 2016. DIGITAL LENDING PARTNERSHIP Headquarters: New York City, New York, United States Year operations launched: 1857; Kabbage partnership Active countries/region of operations: Global. launched in 2016 Partnership is United States only. Headquarters: Santander, Spain Active countries/region of operations: Global. Alternative data utilized: Cloud-based accounting software Partnership is active in United Kingdom only. (e.g., Intuit, Xero), bank account data, and/or POS merchant payment transactions, which reveals indicators such as Alternative data utilized: Kabbage pulls alternative data transaction frequency and volume, seasonal sales, expenses, from bank accounts, merchant acquiring processors, social and customer revenue. The company then analyzes person- networks, e-commerce sites, accounting software, shipping al and business credit history from credit reporting service records (for example, from UPS), and dozens of other private providers, scans public and legal records for past lawsuits or and public sources to gauge the risk and creditworthiness of liens, reviews OSHA11 records for violations, considers the the business seeking the loan. The company’s underwriting health of an applicant’s industry and region, and checks on- engine pulls information such as business revenue, vendor line business reviews from sites like Yelp, Angie’s List, and payments, and tax and accounting data to assign the proper Google Places. line of credit in real time. Target lending market: JPMorgan Chase SME customers. Merchants can optionally link their Facebook, Twitter, and UPS shipping accounts, which may qualify them for fee dis- counts. Facebook business, Yelp, Foursquare, Amazon, and Regions Bank eBay offer business reviews, rankings, and other rich data Website: https://www.regions.com/ on how businesses actually interact with their customers. Type of organization: ESTABLISHED BANK/FUNDATION Similarly, linkages with logistic and e-commerce providers SME DIGITAL LENDING PARTNERSHIP is producing relevant data. Shipping data can reveal how Year operations launched: 1971, Fundation partnership many packages an SME is shipping, how many packages are launched 2015. returned, the longevity of the business, and if the quantity of Headquarters: Birmingham, Alabama, United States packages shipped is increasing or decreasing. Active countries/region of operations: Southern United States. Target lending market: SMEs. Alternative data utilized: Fundation’s technology uses ex- tensive data aggregation, proprietary datasets and intelligent decisioning techniques to predict credit risk and appropri- ately price loans. The power of its offering is in the tools it offers around online applications and data-intensive credit algorithms to partners. Aggregating third-party data in real time, doing a lot of automation, using disparate data sources, 11. The Occupational Safety and Health Administration (OSHA) is the main federal agency in the U.S. charged with assuring safe and healthful working conditions, setting and enforcing standards, and providing training, outreach, education and assistance. 74 ALTERNATIVE DATA TRANSFORMING SME FINANCE Scotiabank Wells Fargo Website: http://www.scotiabank.com/ Website: https://www.wellsfargo.com Type of organization: ESTABLISHED BANK/KABBAGE SME Type of organization: ESTABLISHED BANK/ DIGITAL SME DIGITAL LENDING PARTNERSHIP LOAN AND XERO API LAUNCHES Year operations launched: 1832; Kabbage partnership Year operations launched: 1852; launched FastFlex digital launched in 2016 SME loan and a new digital bank Xero API in 2016. Headquarters: Toronto, Canada Headquarters: San Francisco, California, United States Active countries/region of operations: Global. Active countries/region of operations: United States; Partnership is active in Canada and Mexico only. Global for some businesses. Alternative data utilized: Kabbage pulls alternative data from Alternative data utilized: Created an API so that SMEs can bank accounts, merchant acquiring processors, social net- have their bank account data uploaded directly into the ac- works, e-commerce sites, accounting software, shipping re- counting software provided by Xero. By digitally connecting cords (for example, from UPS), and dozens of other private with the bank, SME customers see their real-time, up-to- and public sources to gauge the risk and creditworthiness of date cash flow each morning on Xero and can receive pay- the business seeking the loan. The company’s underwriting ments faster. This also facilitates credit decision-making and engine pulls information such as  business revenue, vendor allows the bank to better analyze the business as a whole. payments, and tax and accounting data to assign the proper line of credit in real time. In May 2016, Wells Fargo also launched the FastFlex SME loan targeting small business customers with strong cash in- Merchants can optionally link their Facebook, Twitter, and flows and short-term credit needs. The bank built the loan UPS shipping accounts, which may qualify them for fee dis- offering inhouse. The loan amount under this offering ranges counts. Facebook business, Yelp, Foursquare, Amazon, and from US$10,000 to US$35,000 at a competitive interest rate eBay offer business reviews, rankings, and other rich data with one-year terms, and repayments to be made on weekly on how businesses actually interact with their customers. basis. Similarly, linkages with logistic and e-commerce providers is producing relevant data. Shipping data can reveal how many Target lending market: SMEs who are active online. packages an SME is shipping, how many packages are re- turned, the longevity of the business, and if the quantity of packages shipped is increasing or decreasing. Target lending market: SMEs. State Bank of India (SBI) Website: https://sbi.co.in Type of organization: ESTABLISHED BANK/E-COMMERCE PARTNERSHIP Year operations launched: Bank founded in 1806/ e-Commerce partnership in 2016 Headquarters: Mumbai, India Active countries/region of operations: INDIA Alternative data utilized: SBI E-Smart SME uses data an- alytics gathered by Snapdeal to assess the sellers’ cred- it worthiness. It uses these analytics instead of tra- ditional lending based on financial statements such as balance sheet and income tax returns.    See also: http://economictimes.indiatimes.com/article- show/50594157.cms Target lending market: E-commerce SMEs. In partnership with: SME Finance Forum donors