Regulating Innovation for Financial Inclusion: Lessons from Nigeria

Abstract Innovative services such as mobile payments are potentially transformative because they can increase access to financial services, especially in developing countries. However, such innovations can disrupt the financial services ecosystem, prompting regulators to respond in different ways. These regulatory responses often have a significant impact on the success of such innovative services. Using Nigeria's regulatory approach as a case study, this article highlights specific lessons that should inform future attempts at regulating mobile payments.


INTRODUCTION
Domont-Naert identifies that basic financial services include access to a bank account, payment services, credit, insurance and protection against overindebtedness. 1 Where persons cannot access these financial services, they are said to be financially excluded. Financial exclusion is a problem in many developing countries. Banks traditionally provide services through branch networks that are often situated in profitable urban areas. 2 However, a significant populace resides in rural areas where poverty rates are high. Establishing bank branches and automated teller machines (ATMs) to cover Technological innovations that present opportunities for expanding access to financial services have boosted financial inclusion agendas. These innovations cover a range of activities, including offering credit, processing payments, issuing currencies and managing assets, to name a few. 10 One such innovation is mobile payments (m-payments). 11 Belonging to a bouquet of mobile-based financial services, m-payments cover payment transactions in which a mobile device is used to initiate, authorize and confirm an exchange of financial value. 12 M-payments have generated attention in the financial inclusion discourse for several reasons. First, statistics show that the number of mobile phones in circulation exceeds any other technical device that could be used to market, sell or deliver products and services to consumers. 13 In many developing countries, there has been direct implementation of mobile infrastructure rather than a progression from landline to mobile technology, which has facilitated the growth in mobile phone ownership. 14 Being ubiquitous, mobile phones present a practical and cost-effective channel for extending financial services to unbanked persons. 15 Moreover, in these countries, the adoption of mobile-based financial services has been accelerated by the need to address infrastructure gaps that arise when mobile penetration is high, but physical banking infrastructure is deficient. 16 Secondly, certain success stories lend credence to the potential of m-payments to drive financial inclusion. In particular, Kenya's success with its m-payment product "m-pesa" remains a reference point in this regard. Introduced in 2007, m-pesa is a low-cost SMS-based person-to-person money transfer platform that supports the deposit, transfer and withdrawal of funds using mobile phones. 17 Since its introduction, m-pesa has assisted in improving access to payment services in Kenya. Total account ownership 18 in Kenya grew from 42 per cent in 2011 to 81 per cent in 2018. 19 More adults have mobile money accounts (72.9 per cent) in comparison to accounts with traditional institutions (55.7 per cent). 20 M-pesa's success has inspired regulators to prioritise financial inclusion. In 2009, two years after m-pesa's launch, the Central Bank of Nigeria (CBN) designed a regulatory framework for m-payment services in Nigeria. 21 Despite this effort, Nigeria has failed to replicate m-pesa's success. Reflecting on Nigeria's approach to regulating m-payments, this article aims to highlight specific regulatory lessons that can inform future attempts at regulating m-payments. To this end, the article is structured as follows. After this introduction, the next part of the article reviews key aspects of the regulatory framework for m-payments in Nigeria. Building on this, the article then highlights specific lessons that can be extracted from Nigeria's regulatory experience. The next section puts forward alternative considerations that should inform the regulation of m-payments before concluding the discussion.

THE REGULATORY FRAMEWORK FOR MOBILE PAYMENTS IN NIGERIA
In 2010, 46.3 per cent of Nigerian adults were financially excluded. 22 With such numbers, it was unsurprising that, in 2011, the CBN committed to reducing exclusion under the Maya Declaration spearheaded by the Alliance for Financial Inclusion at its Global Policy Forum in Mexico. 23 The Maya Declaration represents "the first global commitment by policymakers from developing and emerging countries to unlock the economic and social 17 MW Buku and MW Meredith "Safaricom and m-pesa in Kenya: Financial inclusion and financial integrity" (2013) 8 Washington Journal of Law, Technology & Arts 375 at 378. 18 The World Bank defines account ownership to include accounts held at a financial institution and those held with a mobile money provider: "Account ownership" (2017), available at: <https://globalfindex.worldbank.org/sites/globalfindex/files/chapters/2017% 20Findex%20full%20report_chapter1.pdf> (last accessed 2 July 2021 potential of the poor through greater financial inclusion". 24 In line with its commitments under the declaration, the CBN launched a National Financial Inclusion Strategy (NFIS) in 2012. 25 A principal goal identified in the NFIS was to increase access to payment services from 21.6 per cent in 2010 to 70 per cent in 2020. 26 The CBN identified m-payments as one of the key drivers in meeting this target. 27 M-payments were singled out because of the high mobile penetration rates at the time. 28 This article divides the key aspects of the CBN's regulatory response into two phases, discussed below.

Phase I: 2009-14
This phase reflects initial attempts to regulate m-payments primarily through the introduction of a licensing regime. Before releasing the NFIS, the CBN issued key regulatory documents applying to m-payments. First, in 2009, it released a document entitled "The Regulatory Framework for Mobile Payments in Nigeria" (the Framework). 29 In 2014, it released a second document entitled "Guidelines on Mobile Money Services in Nigeria" (the Guidelines), which updated the Framework. 30 The Guidelines outline the three objectives informing the CBN's intervention. First, the intervention aimed to "ensure a structured and orderly development of mobile payment services in Nigeria, with clear definition of various participants and their expected roles and responsibilities". 31 Secondly, it aimed to specify the minimum technical and business requirements for the various participants in the recognized m-payments industry. 32 Thirdly, it sought to promote the safety and effectiveness of m-payment services to enhance user confidence. 33 Under the Guidelines, only licensed entities designated as "mobile money operators" (MMOs) may provide m-payment services. 34 The Guidelines permit two business models that MMOs can adopt in providing the services: the bank-led model and the non-bank led model. 35 The bank-led model is one in which the service provider is a bank, acting either alone or in a consortium with other banks. In this model, banks may choose to partner with other approved organizations but remain principally responsible for delivering 24 CBN "National financial inclusion strategy", above at note 22 at 23. 25  the services. 36 In the non-bank model, a corporate organization, other than a bank, may be licensed to deliver m-payments. While not defining which organizations qualify to lead a non-bank model, the Guidelines expressly exclude mobile network operators (MNOs) from the model's scope. 37 The CBN justified this exclusion on the basis that MNOs would pose a threat to the stability of the financial system. This is because MNOs lack the experience of implementing financial regulations and would expose the financial system to systemic risks. 38 The Guidelines limited MNOs' role to providing the telecommunication network infrastructure for use by MMOs. 39 To this end, MNOs are required to implement a secure communications channel that is compliant with the technology standard stipulated in the Guidelines. 40 To encourage competition, MNOs are precluded from giving preferential treatment to any specific MMO in terms of price and traffic. 41 Furthermore, MNOs must ensure that their subscribers are free to use any m-payment service of their choice. 42 Reiterating their exclusion from providing payment services, the Guidelines enjoin MNOs not to receive deposits from the public except for their subscribers' airtime billing. 43 They are also precluded from allowing the use of prepaid airtime value loaded by their subscribers to transfer monetary value or for payment purposes. 44 Although the CBN released the Framework in 2009, it only began to receive applications for licences in 2010. By 2014, the CBN had issued licences to 21 MMOs. 45 During this period, some banks had begun to include m-payment services as part of the suite of services available on their mobile banking platforms. Many non-banks that received a licence were unable to commence operations due to limited funds to build the required infrastructure and agent networks. 46 The few non-bank platforms that were operational 47 did not gain 36 Id, para 5.0(a). 37 Id, para 5.0(b). 38 GSMA "What could we learn from Nigeria barring MNOs from participating in the mobile money market?" (29 April 2013), available at: <https://www.gsma.com/mobil efordevelopment/region/sub-saharan-africa-region/what-could-we-learn-from-nigeriabarring-mnos-from-participating-in-the-mobile-money-market/> (last accessed 2 July 2021). 39 The Guidelines, para 8.4. See also CBN "Regulatory framework", above at note 21, para 2. the traction comparable to counterparts like m-pesa. This was probably because they were not recognizable brands that could inspire consumer trust and they did not have extensive agent networks.

Phase II: 2015-2020
Despite the number of licences issued in the first phase, the m-payment market failed to record significant success in Nigeria. A 2016 survey revealed that 76 per cent of Nigerians were unfamiliar with mobile money and 98.7 per cent had never registered for or used a mobile money service. 48 Owing to this failure, the CBN began to embrace regulatory shifts, most of which focused on ensuring that MMOs could access agent networks of other stakeholders, like MNOs. The CBN's focus is explained by its admission that, "agent networks present an opportunity to service people in areas that lack bank branches or other physical financial access points like ATMs. Consequently, a functional agent network is imperative for extending financial services to the unbanked. However, deficit of fixed location agents has been a challenge". 49 In comparison to banks and other firms permitted to provide m-payments, MNOs have a well-dispersed network of existing outlets across Nigeria. In 2015, a geospatial mapping survey of MNO access points captured about 8,533 operational outlets in Nigeria's 36 states and Federal Capital Territory. 50  whom they will be responsible. 53 The scope of banking-related activities that super-agents or agents can undertake is defined under the CBN's Guidelines for the Regulation of Agent Banking and Agent Banking Relationships in Nigeria. 54 Functions with which agents can assist include accepting cash deposits and withdrawals, bills and salaries, and local funds transfer. 55 Super-agents can only use their platform to manage their agents' activities and are precluded from holding electronic money. 56 Only licensed financial institutions remain permitted to provide and operate mobile money platforms and to hold electronic money. 57 The CBN, in collaboration with relevant stakeholders, also launched the Shared Agent Network Expansion Facility programme to encourage the development and sharing of agent networks. 58 Under this programme, participating institutions can share agents in rolling out financial services. This programme and the Super Agents licensing regime viewed MNOs as "distribution actors", 59 whose agent networks could be exploited. They remained excluded from taking the lead in providing services like m-payments.
In 2018, the CBN revised the NFIS. 60 A key issue recognized in the implementation of the 2012 NFIS was that mobile money had failed to take off, owing to restrictive policies. 61 The revised NFIS highlighted overarching policy principles that would inform its implementation. The first principle hinges on the adoption of an appropriate risk-based regulatory level playing field that ensures that the same set of regulatory requirements and conditions applies to all potential providers of a particular service, regardless of their background or type of operation. 62 The second principle focuses on encouraging relevant actors to play to their core areas of strength to achieve high impact.
These principles appear to have set the tone for the policy change heralded by the introduction of the Guidelines for the Licensing and Regulation of Payment Service Banks in Nigeria (PSB Guidelines). 63  20PAYMENT%20BANK.pdf> (last accessed 2 July 2021). These were revised in August 2020; see CBN "Guidelines for licensing and regulation of payment service banks in Nigeria" (August 2020), available at: <https://www.cbn.gov.ng/Out/2020/CCD/APPROVED%20 REVIEWED%20GUIDELINES%20FOR%20LICENSING%20AND%20REGULATION%20OF% provide the framework for licensing niche banking institutions, called payment service banks (PSBs). The main objective behind setting up PSBs is to "enhance financial inclusion in rural areas by increasing access to deposit products and payment / remittance services … through high-volume low value transactions in a secured technology-driven environment". 64 PSBs are permitted to maintain savings accounts and to accept deposits. They can also provide payment and remittance services and may operate an electronic purse. 65 They cannot, however, grant loans, advances or guarantees. 66 In what appears to be a significant departure from the previous policy stance, the PSB Guidelines permit MNOs, through subsidiaries, to register as PSBs. Under this framework, MNOs may provide payment services if they meet the licensing requirements and are granted approval-in-principle. 67 In September 2019, the CBN issued approvals-in-principle to three entities, two of which are controlled by MNOs. 68 While it is still too early to assess whether the inclusion of MNOs as eligible promoters of PSBs will drive the m-payment market, Nigeria's experience so far and this policy U-turn provide some lessons that can inform future attempts at regulating m-payments.

MATTERS ARISING: WHAT LESSONS CAN WE LEARN?
The participation of mobile network operators may be decisive in the successful roll-out of mobile payments The CBN's evolving position on MNOs directly providing m-payments raises questions about the MNOs' role in driving the market. This article takes the position that the CBN's initial exclusion of MNOs from direct participation was detrimental to the growth of the m-payments market. There are several reasons for this view. First, it is supported by inferences from empirical research carried out by Evans and Pircho. 69 Their study focused on discovering why m-payments succeeded in some countries and failed in others. The study identified eight countries in which m-payments recorded explosive contd 20 PAYMENT%20SERVICE%20BANKS%20IN%20NIGERIA-27AUG2020.pdf> (last accessed 2 July 2021). 64  growth. 70 One common characteristic in these countries, save for one, 71 was that there was no regulatory restriction on which parties could provide m-payments. In identifying the factors that could stimulate success, the study asserts that "the regulatory framework adopted by the government, in particular, the extent to which regulations restrict potential players, in particular mobile network operators (MNOs), from operating mobile money schemes … could facilitate or restrain success". 72 Among 22 countries investigated by Evans and Pircho, there was only one success story that was not driven by an MNO-led scheme. This service, bKash, is sponsored by the BRAC Bank in Bangladesh. The service is reportedly successful because it is supported by the MNOs that account for most subscribers in the country. 73 However, the service has been mostly useful in providing a platform that allows people to pay bills. Unlike with services that are MNO-led, it has not been successful as a platform for person-to-person transfers. 74 Pakistan also has a successful bank-led platform, but its experience has been unique. In Pakistan, MNOs are not allowed to provide m-payments direct. To circumvent this, Telenor Pakistan, Pakistan's largest telecoms operator, acquired a 51 per cent stake in Tameer Bank to launch its m-payment service, Easypaisa. 75 Easypaisa enjoys an extensive agent network backed by Telenor Pakistan's distribution networks. 76 Evans and Pircho's study appears to confirm that the key to unlocking m-payments rests with flexible regulations that permit direct participation by stakeholders such as MNOs. Their conclusions are supported by a 2017 study carried out by Riley and Kulathunga, investigating four jurisdictions that successfully leveraged electronic money and digital financial services in driving financial inclusion. 77 They found that the involvement of the private sector and non-bank entities, which were supported by flexibly designed policies, positively impacted financial inclusion statistics. 78 MNOs are singled out as important stakeholders for several reasons. First, they have a communications network that allows for transactions to be performed in real-time. 79 Secondly, they operate the subscriber identity module (SIM) cards that provide the technology to support SMS-based payment services such as m-pesa. Thirdly, they have easily recognizable brands and a vast retail distribution network of airtime sellers, even in rural areas. 80 They are also well-experienced in running high-volume, real-time prepaid platforms to a high standard of availability and reliability. 81 The regulatory shifts occurring in several countries, including Nigeria, also support the conclusion that MNOs' participation is critical in driving the m-payments market. Evans and Pircho's study identified eight countries in which m-payments have failed to ignite. 82 The study found that all eight (including Nigeria) adopted a heavy-handed approach to regulating m-payments. Of the eight, seven required that banks take the lead or significantly participate in providing the service. 83 Since the study was published, regulatory shifts have occurred in at least three countries: Ghana, India and Nigeria. 84 In Nigeria and India, as noted below, central bank authorities have moved from policies prohibiting the direct involvement of MNOs, to policies allowing MNOs to participate in the provision of m-payments through niche financial institutions (payments banks).
In Ghana, the 2008 Guidelines for Branchless Banking favoured a bank-led business model for the provision of m-payments and required that at least three banks be involved in any service. 85 This approach was described as "many to many" and aimed to encourage interoperability, which would allow the m-payments market to take off swiftly. 86 MNOs were excluded from applying for licences independently and were required to partner with banks. The m-payments market did not take off as expected. Research by the Consultative Group to Assist the Poor (CGAP) found that banks had little incentive to make significant investments under these arrangements. 87 MNOs made most of the investments and felt disadvantaged because the regulations did not permit them to take an independent lead in providing the service. 88 The failure to record success prompted regulatory reform, which saw the passing of new guidelines for e-money issuers in 2015. The 2015 guidelines 80  permit MNOs to apply to Ghana's Central Bank for licences. 89 These regulatory shifts lend credence to the proposition that MNOs are critical in driving the m-payments market.
Clues from Kenya: Flexible regulatory approaches may be more appropriate for innovative financial services like mobile payments Nigeria's experience also raises questions about the impact of a chosen regulatory approach in the growth of innovative financial services like m-payments. This issue is significant, because a regulator's approach may provide either the right conditions that encourage innovation or may hinder its growth by compounding the risk already inherent in the acceptance of a novel product. 90 Further light can be shed on this issue by juxtaposing the approaches of the CBN and Central Bank of Kenya (CBK) in Nigeria and Kenya, respectively. Safaricom, an MNO, provides Kenya's leading m-payment product, m-pesa. When m-pesa was launched in 2007, there were no regulations dedicated to m-payments. As financial inclusion was a policy priority for the CBK, it was interested in the success of m-pesa. 91 The regulatory approach involved continuous dialogue between the CBK and Safaricom before the product's launch. When Safaricom submitted its proposal, the CBK conducted an internal review, which focused on clarifying specific areas of concern. 92 First, the CBK was concerned about the legal status of m-pesa: it needed to decide whether to classify it as a banking business. Secondly, it was concerned about the money laundering risks introduced by the product. Thirdly, the CBK sought to understand the operational risks associated with the service.
Following legal advice, the CBK reached several conclusions that informed its regulatory stance. 93 First, it decided that m-pesa was not a banking service as defined under Kenya's Banking Act. 94 This was because the cash exchanged for electronic value was not repaid on demand and remained in the 89  customer's control. 95 Secondly, it concluded that there was no credit risk for customers or Safaricom, because m-pesa agents were required to make upfront deposits of cash in m-pesa accounts held by local banks. 96 Thirdly, the CBK established that there was no intermediation: 97 customer funds were not lent in the pursuit of other business, interest or income, and all funds were held in a trust account and could not be accessed by Safaricom to fund other parts of its business. 98 Fourthly, the CBK found that m-pesa was developed with anti-money laundering (AML) measures in mind. There were functions supporting the generation of electronic trails and suspicious transaction monitoring. Transaction caps were also set on individual and aggregate daily transactions and international remittances. 99 Finally, the CBK concluded that m-pesa's operational risk was minimized as there was end-to-end encryption of the SIM card to ensure security and live back-up. The CBK requested that Safaricom undertake comprehensive technical assessments carried out by Consult Hyperion 100 to evaluate the operational risks of the m-pesa platform. 101 The service passed Consult Hyperion's tests for operational capacity. 102 It was found that there were reporting and monitoring mechanisms that ensured that the CBK could request information concerning the firm's audit trail, AML procedures, liquidity management and clearing / settlement. 103 The CBK also held consultations with the Communications Authority of Kenya, Safaricom's primary regulator. 104 These consultations revealed that the authority considered m-pesa to be a value-added service that Safaricom was licensed to offer. 105 Based on these findings, m-pesa was not regulated as a financial service. The CBK concluded that m-pesa had adequate controls in areas that could affect financial stability and issued it with a letter of no objection. 106 95 AFI "Enabling mobile money", above at note 92 at 4. 96 Ibid. 97 This is the process by which banks take in funds from a depositor at low-interest rates and lend them out at higher interest rates to make a profit. M-pesa's success has partly been attributed to the liberal regulatory approach of the CBK, 107 which has been described as a "test and learn" method that is implemented on a case-by-case basis. 108 With this approach, non-bank providers can benefit from regulatory forbearance, like the "letter of no objection" granted by the CBK or restricted licences or special charters. 109 In such circumstances: "In return for the regulator's 'clarification' that the FinTech firm's activity is outside the scope of certain rules which are viewed as unnecessary or inappropriate under the circumstances or in the specific context, the no-action letter or restricted license may be supplemented with conditions seeking to ensure that even if certain rules do not apply, the principles underlying the regulation are still upheld. The practical effect of forbearance through no-action letters, restricted licensing, or special charters is that of partial exemptions or dispensation within a broader regulatory framework." 110 The advantage of this approach is that it encourages communication between regulators and innovators. Regulators can study business models and risk assessments and request clarification where needed. 111 This allows regulators to obtain sufficient data and experience to inform their regulatory response. 112 However, the main drawback with this approach is that it is more suitable where the number of firms requesting exemptions is small. Where more firms request bespoke exemptions, this will increase costs and put a strain on regulatory capacity. 113 It may also be challenging to ensure that equal treatment is extended to each participant. 114 This can lead to unintended consequences. For instance, from a competition perspective, Kenya's approach provided Safaricom with a dominant market position. Safaricom's dominance attracted criticism that forced regulators to call for the implementation of interoperability between m-pesa and other mobile money services. 115 Such dominance can negatively impact competition, leaving consumers with limited choices, since providers have little incentive to improve the quality of their services or preserve their reputation. 116 It can also lead to the creation of "too big to fail" providers, inviting serious consequences. The failure of a large provider like Safaricom can lead to system-wide disruption in the economy and can have serious reputational consequences for regulators. 117 Such disruptions can also discourage potential users of similar services, which can jeopardize financial inclusion gains.
Zetzsche et al also point out that this approach may lead to long-term uncertainty for businesses. 118 This suggests that a flexible case-by-case assessment is more appropriate in the early stages of regulating a new product. This approach in the early stages of regulation allows regulators to gather the information and experience required to issue rules that will apply to future innovators. Again, Kenya's experience fits with this. Despite its success, m-pesa highlighted a gap in the regulation of payment services in Kenya. 119 Consequently, the National Payment Systems Regulations (NPSR) was passed in 2014. The NPSR introduced new regulatory rules for the payment sector and identified the CBK as the primary supervisory authority for payment service providers (PSPs). 120 The NPSR set out basic e-money rules and require interested firms to apply for authorization or registration. 121 MNOs may be designated as PSPs or e-money issuers. 122 Unlike Kenya, Nigeria adopted a less flexible approach to regulating m-payments. The CBN released the Framework document in 2009 before any m-payment service was launched and only began to receive licence applications in 2010. Out of about 40 applicants, only 18 received approvals-in-principle, with a requirement to build m-payment platforms and run pilots within three months. 123 Certain problems are associated with the CBN's approach. and issues" (discussion paper, 2014-2) at 2, available at: <https://www.bankofcanada. ca/wp-content/uploads/2014/04/dp2014-2.pdf> (last accessed 2 July 2021). 118 Zetzsche et al "Regulating a revolution", above at note 108 at 63-64. 119 AFI "Enabling mobile money", above at note 92 at 6. 120 NPSR, reg 30. Kenya's National Payment System Act (2011), sec 2 defines a "payment service provider" as: "(i) a person, company or organisation acting as provider in relation to sending, receiving, storing or processing of payments or the provision of other services in relation to payment services through any electronic system; (ii) a person, company or organisation which owns, possesses, operates, manages or controls a public switched network for the provision of payment services; or (iii) any other person, company or organisation that processes or stores data on behalf of such payment service providers or users of such payment services." 121 NPSR, reg 4. First, the 2009 Framework contained prescriptive requirements dictating the acceptable business models, technology, methods through which m-payments could be carried out, and operational arrangements for delivering m-payments. By designing regulations before any m-payment service was launched and before considering any approvals-in-principle, the CBN did not have the opportunity to understand the service before it issued regulations. Secondly, the CBN failed to take advantage of the chance to gather relevant information during the mandated pilots. The CBN required applicants to build and pilot m-payment platforms within three months. This pilot period was short, and only two applicants met the deadline. 124 This contrasts with Kenya, where m-pesa was piloted with the CBK's support two years before its launch. 125 This gave sufficient time for Safaricom to adjust its product offering to reflect market needs and also gave the CBK more time to gather information on the product. 126 The CBN's short mandatory pilot period arguably failed to give applicants enough time to adjust their product offering. It also does not appear that the pilots enabled the CBN to gather further information that would influence regulations. This is because the regulatory framework was not updated until four years after the pilots.
The contrast in the experiences of Kenya and Nigeria highlights the dilemma that authorities face in deciding what regulatory approach is more appropriate for innovative products. No consensus exists on the most suitable regulatory strategy. While some writers suggest full regulation resembling prudential banking regulations, others call for a "tailored regime" and warn against "unnecessary or over-reactive" regulation. 127 Although Nigeria's early adoption of full regulations can be interpreted as heavy-handed, it can be justified on the basis that maintaining the safety of the financial system is at the core of the mandate of central banks. 128 However, the problem rests in the inability to draw a balance between protecting the financial system and encouraging innovation, which in turn can increase inclusion. Since innovation can positively impact financial inclusion, it is reasonable to embrace a flexible regulatory approach like Kenya's, which can adapt to changes in the financial services landscape. 129 Regulators can implement Kenya's test and learn approach in ways that best fit their jurisdictions. An increasingly popular and more structured option is the use of regulatory sandboxes. 130 A regulatory sandbox is a "safe space in which businesses can test innovative products, services, business models and delivery mechanisms without immediately incurring all the normal regulatory consequences of engaging in the activity in question". 131 With standardized eligibility requirements, sandboxes aim to encourage competitive innovation while enhancing the open and transparent exchange of information between regulators and innovators. 132 Such a flow of information can inform future regulatory policies. In our context, sandboxes can be useful for emphasizing financial inclusion objectives. Where this is a policy priority, regulators can implement thematic sandboxes where preference is given to innovation that supports inclusion. 133 As a recent approach to regulating financial services innovation, there is not yet enough empirical evidence of the impact of sandboxes. 134 However, early research suggests that, while sandboxes can contribute to developing evidencebased policy, they are insufficient in themselves in promoting innovation or financial inclusion. 135 Hence, regulators must engage in implementing broader reforms that support innovation, regulatory capacity, market engagement and financial inclusion. 136 As sandboxes are not a "one-size-fits-all" solution, some researchers call for a focus on broader alternatives that reflect the priorities and capacity of regulators, as well as encouraging the general development of  innovation ecosystems. 137 These alternatives could focus on providing dedicated spaces (such as innovation hubs) or institutions (such as innovation offices) aimed at supporting innovators.
Niche financial institutions may present an attractive compromise but do not guarantee results As indicated earlier, the latest regulatory update in Nigeria has seen the introduction of niche financial institutions (PSBs), which are permitted to provide m-payments. Under the new framework, MNOs through their subsidiaries can participate directly in providing m-payments. This appears to be a compromise between two policy positions: excluding MNOs from providing the service; and allowing them to participate directly, but within the regulatory control of the CBN. The idea of PSBs is not novel and Nigeria may have been inspired by India. Like Nigeria, India's initial response to regulating m-payments was to exclude MNOs from providing the service. Only licensed banks with a physical presence in India were initially permitted to offer m-payments. 138 These services were also restricted to existing bank customers and / or credit and debit cardholders. 139 Justifying its initial strict bank-led approach, the then deputy governor of the Reserve Bank of India (RBI) declared that the RBI believed that, "sustainable financial inclusion is achievable only through mainstream financial institutions, ie banks". 140 He further explained that, "[i]n India, we have adopted a bank-led model for financial inclusion, which seeks to leverage on technology … Our experience shows that the goal of financial inclusion is better served through mainstream banking institutions as only they have the ability to offer the suite of products required to bring in effective / meaningful financial inclusion". 141 In 2014, the RBI adjusted its stance by opening the playing-field to nonbanks. 142 However, it held onto its initial belief that banking institutions were better placed to drive inclusion. Hence, the RBI only permitted nonbanks to provide payment services if they registered as newly created banking institutions (payments banks) offering a range of financial services. Accordingly, it issued Guidelines for the Licensing of Payments Banks (Licensing Guidelines). 143 The RBI stated that the primary objective for introducing payments banks was to further financial inclusion. 144 MNOs, supermarket chains, non-banking finance companies and public sector companies are among the classes of persons eligible to apply for licences. 145 The regulatory change provided leeway for MNOs to participate in the provision of m-payments. As in Nigeria, payments banks may provide savings accounts and payment / remittance services, but cannot offer credit. 146 The justification for payments banks is that they offer a broad suite of services beyond m-payments. Their provision of additional services reflects the RBI's belief that a superior model for improving inclusion is one that supports additional banking services alongside m-payments. 147 While this is an important consideration, if there is no demand for the additional services, then the introduction of payments banks may be strategically unwise. This is because, for innovation to succeed, it should be designed to respond to an unmet need expressed by consumers and should not be based on assumptions of what consumers want. 148 Some examples buttress this point. M-pesa is partly successful because the product's design addressed the existing need for low cost person-to-person transfers. 149 Although the initial product proposal focused on microfinance transactions, consumer feedback during the pilot prompted changes that made the product functionally relevant. 150 Reflecting the market need at its launch, Safaricom focused on providing a convenient and costeffective domestic platform for sending money from urban areas to rural communities. 151 This can be contrasted with unsuccessful attempts in other jurisdictions. In South Africa, for instance, a product called WIZZIT was introduced to provide a full suite of banking services to the unbanked. Besides stiff regulatory roadblocks faced by WIZZIT's founders, a Harvard study suggests that one of the reasons behind the product's failure was that it was launched without much consideration of the market demand for the services offered. 152 So far, payments banks have not made a substantial impact in India. While 11 in-principle approvals were granted to applicants in 2015, by 2018 only four of those 11 remained operational. 153 These banks registered weak performances, incurring net losses in the financial years 2016-17 and 2017-18. 154 Their lacklustre performance has been attributed to limited revenue streams. 155 This is a direct consequence of stringent regulations that dictate the business model that they can adopt. 156 Although registered as banks, they are prohibited from engaging in any lending activity and cannot make any profit from interest. The revenue margins from deposits are also small as there is a cap on the deposits that they can accept. 157 At the same time, they must offer attractive interest rates on deposits if they wish to compete with traditional banks. 158 Although it is too early to assess fully the impact of niche institutions like payments banks, two points must be made. First, while it is important that excluded persons can access a wide range of financial products in the longterm, it may be more sustainable to allow providers to develop products organically that reflect market needs. Hence, user-driven extensions of essential payment services may be more viable in the long term. Understanding the needs of the underserved and building their trust and familiarity with basic digital financial services is likely to convert them into account-holders in the long-term. 159 M-pesa continues to stay relevant because it constantly adapted and extended its functionalities based on customer feedback and changing market trends. 160 From focusing on basic services like facilitating person-to-person payments, airtime purchases and cash withdrawals from agents, m-pesa has expanded its services over the years to include transfers to and from regular bank accounts, international remittances and microinsurance. 161 It also went on to partner with the Commercial Bank of Africa to offer M-Shwari, a micro-credit service. 162 Secondly, India's experience raises questions about whether niche banking institutions will encourage MNO involvement in driving the m-payments market. India's experience may suggest that such institutions can disincentivize MNOs from investing. MNOs are not financial institutions but will be subject to stringent banking regulations if they register as niche banking institutions. With restrictions on their activities limiting the profits that they can make, they may decide that investing in the market is not entirely justified.
Regulators must carefully consider which mobile payment models will best achieve identified policy goals M-payments can be provided using several business models. There are divergent views on which model is best suited for driving the market. For instance, Rajan argues that the best model for m-payments is one where MNOs only act as intermediaries between consumers and financial institutions. 163 This view is justified on the basis that financial regulations are onerous and limiting the role of MNOs to intermediaries will ensure that the cost of extending financial regulations to MNOs is avoided. 164 Moreover, while innovations like m-payments can increase financial inclusion, they are disruptive and can introduce negative externalities. 165 These externalities may include prudential risks, which can negatively affect the safety of the financial system. Non-banking institutions have not traditionally been saddled with managing financial transactions and, therefore, have a limited understanding of the prudential risks involved. Their limited experience further suggests that they will also have limited risk management capabilities. 166 Arguments in favour of bank-led models are often justified on the basis of these points.
In contrast, commentators such as Alexandre et al argue that, in jurisdictions where banks cannot profitably serve a significant number of people, regulators should allow a broader range of participants, which face different cost structures and economic incentives, to contest the market directly. 167 Supporting this view, Porteous argues that, given the weakness of the retail banking sector in many developing countries, it is necessary that non-bank 161 Id at 1208; Hughes and Lonie "M-pesa: mobile money", above at note 125 at 78. players, particularly MNOs with their strong retail brands and established networks, be able to issue e-money. 168 He further argues that, even if non-banks decide not to invest, the threat of possible entry may galvanize a response from banks. 169 Under this approach, non-bank players like MNOs may be permitted to provide m-payments directly in different ways. One method may allow MNOs to acquire direct licences to provide the service without any substantial changes to their structure. This was the case with Kenya's m-pesa. Another method may require that they participate through niche banking institutions, as is the case in India and Nigeria.
While banks and MNOs remain critical to providing fund settlement and mobile network facilities respectively, m-payments can be provided by other electronic money firms that are neither banks nor MNOs. 170 The involvement of such firms promises a more competitive market that will leave consumers with increased choices. While promising, this model can face certain difficulties. First, initial investment costs will be high, as such providers will need to enter separate agreements with participating banks and MNOs to run their service. Secondly, where such providers are new, they may not be popular enough to inspire consumer trust, and many excluded persons may refrain from embracing their services. Thirdly, such providers may have limited agent networks due to the high costs of establishing them. If this is the case, they will struggle to achieve the coverage that MNOs and, to a lesser extent, banks enjoy. This will mean that, in many developing countries, they may not gain traction outside urban areas. Finally, owing to the scale of the services that they provide, MNOs and banks are well-experienced in operations management, which other providers may not be. Where providers lack such capacity, their inexperience can lead to operational challenges, which may erode their business. 171 Some writers argue for hybrid models involving alliances between the two crucial parties in the m-payments transaction process: banks and MNOs. This model is thought to represent the most beneficial model for all parties 168  involved, because it allows them to retain their core functions. 172 Where collaboration occurs, banks will leverage their experience in settling payments, while MNOs will leverage the extensive reach of their agent networks. While this model has its advantages, collaboration between key stakeholders can be challenging to achieve. For example, despite m-pesa's success, Safaricom's 173 early attempts to collaborate with a bank (Equity Bank) in introducing another platform, M-kesho, were unsuccessful. This failure was partly attributed to the difficulty associated with managing complex issues, like the division of responsibilities and the sharing of revenue, which led to friction. 174 Despite these challenges, this model may be the most sustainable in the long term since it combines the advantages of the bank-led and MNO-led models. A model that allows parties to focus on their core competencies will probably make the most economic sense to competing participants. Hence, in the long run, it is anticipated that participants may voluntarily choose to collaborate in offering m-payments. As Alexander et al point out: "Product development will remain the preserve of banks, as telecom companies have no expertise and likely no desire to develop financial products beyond basic transactional services. This difference in aspirations and comparative advantages should establish a strong basis for partnerships between mobile operators that will be motivated by the growth in transaction volume and banks that will be more interested in accessing float and cross-selling and up-selling products to clients". 175 While agreeing that a hybrid may be the most useful model, it is submitted that such collaborative partnerships should occur organically, based on voluntary terms agreed between participants. If regulators mandate hybrids, participants may be reluctant to invest if they are not ready to make the necessary commitments. Affirming this argument, in many countries where providing m-payments is not restricted to banks, MNOs have chosen to partner with them. 176 Although the Kenyan Bankers Association initially resisted m-pesa, many Kenyan banks incorporated m-pesa with their own core applications 172 L Chaix and D Torre "Four models for mobile payments" (2011), available at: <https:// www.researchgate.net/profile/Dominique_Torre/publication/267917243_Four_models_for_ mobile_payments/links/547486d00cf2778985abe525/Four-models-for-mobile-payments. pdf?origin=publication_detail> (last accessed 3 July 2021 to reduce their operational costs, setting the stage for further collaboration.
Since 2010, Safaricom, the MNO behind m-pesa, has formed partnerships with several banks to offer other products, like savings accounts, micro-credit and insurance. 177 In deciding what model to promote, regulators must consider how the merits and demerits of the different arrangements highlighted above can impact identified policy objectives. Porteous argues that enabling regulatory environments should allow for the development of business models that meet a defined policy objective. 178 Hence, if a regulator's objective is to increase financial inclusion using m-payments, then it ought to support the models that can make this possible. If a significant percentage of a country's populace cannot access formal financial services due to limited bank branches or service channels, then regulators should embrace models that encourage the participation of actors that have good agent networks in underserved areas. This suggests that MNO-led models or hybrid models will probably record transformative success in comparison to bank-led models. This is because, if a primary reason for exclusion is poor bank branch penetration, addressing this problem by introducing a bank-centric solution only leads back to the origin of the problem. If regulation places banks at the centre of driving the m-payments market, it indirectly makes a relationship with a bank a prerequisite for m-payments, which in effect shuts out already excluded consumers.
In many countries where mobile money services have contributed to increasing inclusion, regulations have permitted flexible models that support the participation of the stakeholders with extensive agent networks. Apart from Kenya, another example may be seen in Bangladesh. Bangladesh has taken advantage of the extensive reach of its post office network to encourage models that allow for the delivery of digital financial services. The post office department has an extensive network of 9,886 post offices across the country, which has proved vital in providing a mobile money order service. 179 This department provides a mobile money order service in partnership with the second largest MNO in Bangladesh (Banglalink Mobile Company), which has an extensive network coverage area. This partnership enables the service to be accessible at numerous postal outlets, including in areas with little or no internet connectivity.

LOOKING AHEAD: THINKING BEYOND THE PLAYERS AND FOCUSING ON THE REGULATORY RISKS
M-payments raise several regulatory concerns, which may include the potential abuse of the financial system, data protection risks and consumer protection challenges. For regulators wishing to leverage m-payments in driving financial inclusion, the key consideration should be how they can design appropriate responses that address these risks. This is because the positive impact of any form of innovation, including m-payments, will only remain credible in the long term if regulators can identify and address the specific risks that attract negative externalities. Accordingly, regulatory resources are best channelled towards analysing trends in the m-payments market to identify and understand risks, and design safeguards to address them. 181 For instance, where regulators permit non-banking institutions to provide m-payments, those institutions are unlikely to have the sophisticated experience required for asset and liquidity management. This inexperience can introduce systemic risks, requiring regulators to adopt mechanisms that can lessen the risks. In Kenya for example, the NPSR introduce a core capital requirement for authorized PSPs. 182 PSPs are also prohibited from engaging in any lending or investment activity unless expressly permitted under the NPSR. 183 Where a PSP is involved in other unrelated ventures, it must keep its payment service in a separate business unit, and must maintain a separate management structure and books of account. 184 PSPs are consequently mandated to create a trust to cover consumer funds. 185 They are also obliged to adopt appropriate risk mitigation strategies that ensure that consumer funds are placed in commercial banks and diversified. 186 Similarly, in India, payments banks are required to invest 75 per cent of their demand deposit balances in government securities / treasury bills with a maturity of up to one year. 187 They must also hold a maximum of 25 per cent in current and time / fixed deposits with other scheduled commercial banks. 188 All other financial and non-financial activities of a bank's promoters must be ring-fenced and should not be mingled with the banking and financial services business of the payments bank. 189 To provide a buffer against operational risks, the RBI requires that a minimum paid-up equity capital of one hundred crores (approximately GBP 9.6 million) be maintained. 190 Another critical concern for regulators is ensuring that innovation does not leave the financial system open to abuse. To achieve this, regulators must impose arrangements that seek to maintain financial integrity. For instance, the potential misuse of the financial system may be mitigated by requiring providers to adopt customer registration and verification measures that ensure that persons using the services can be identified. In reality, it can be difficult to balance such arrangements with financial inclusion policies. This is because regulators and regulated institutions are often cautious about breaching international standards aimed at combating abuse of the financial system. This can lead to inflexible and inappropriate measures that may prevent people from accessing formal financial services. 191 For instance, in many developing countries, some persons find themselves financially excluded because they cannot satisfy regulatory requirements, such as those mandated for customer identification and verification purposes. 192 To address this, the Financial Action Taskforce (FATF), the lead inter-governmental body that sets standards aimed at combating money laundering and terrorist financing, has released several guidance documents to support regulators. 193 Central to the FATF's advice is a call for money laundering and terrorist financing regulators to apply a risk-based approach. 194 This means that regulators should develop flexible and proportionate responses that address actual risks flagged up in their risk assessment. The understanding is that, if regulators take advantage of the flexibility allowed by the risk-based approach, they can tailor measures that are attuned to their jurisdictions and will not negatively impact financial inclusion.
The FATF also requires that financial institutions develop internal policies that address money laundering and terrorist financing risks 195 and encourage the inclusion of AML / CFT 196 risk assessments in the development of new products. 197 The m-pesa development process exemplifies adherence to this. 198 As discussed earlier, the CBK satisfied itself that the product was designed to mitigate money laundering and terrorist financing risks and that Safaricom's internal AML / CFT policy was sufficient. This was reflected in functionalities that supported transaction caps, 199 suspicious transaction reporting 200 and the generation of electronic transactional trails. 201 Practical implementation of the FATF guidance will require that regulators also implement other broader reforms. For example, to ensure that customer identification and verification processes are not exclusionary, regulators must develop national identification systems, which will make it easier for excluded persons to satisfy such requirements. Beyond Kenya's regulatory flexibility, m-pesa's success has also been attributed to the fact that Kenya already had a workable national identification system when it was launched. This helped to lessen the customer registration and verification requirements needed to open an m-pesa account. 202 De Koker and Jentzsch point out that, while adherence to the FATF standards promises increased transparency in the financial system, it can introduce other considerations from a privacy perspective. 203 AML procedures require that personal information about customers as well as their transaction trails are collected and stored. In developing countries with weak data protection frameworks, this can be problematic. The potential for unremedied data breaches and the likelihood that personal and transaction information can be misused, even by government authorities, may discourage people from transacting within the formal financial sector. 204 To alleviate these concerns, regulators and lawmakers must design robust privacy and data protection laws.
Excluded persons are usually the target of innovative products that seek to drive inclusion. Because these persons are often vulnerable due to their circumstances, regulators must design measures that ensure that they are no worse off in adopting such services. 205 Such efforts will cover issues such as protecting consumer funds, which can be addressed by the prudential measures discussed above. These efforts should also address other consumer protection concerns that may lead to personal consumer detriment. 206 With m-payments, for instance, consumers may be subjected to unfair commercial practices, such as the use of unfair contract terms. M-payment services will be offered to consumers on standard form contracts, which will be available on a take it or leave it basis. Such contracts are prone to abuse, as they are one-sided and cannot be negotiated. This problem is exacerbated in many developing countries where illiteracy levels are high, making it almost impossible for consumers to understand the consequences of the terms on which they contract.
Tackling the use of unfair commercial practices will require wider legislative reforms that are broadly drafted to accommodate technological innovation. As stated earlier, m-pesa's success highlighted regulatory gaps that spurred legislative reforms in several areas, including consumer protection frameworks. One of the aims of the Kenyan Consumer Protection Act, which was passed in 2021, was to protect "consumers from all forms and means of unconscionable, unfair, unreasonable, unjust or otherwise improper trade practices including deceptive, misleading, unfair or fraudulent conduct". 207 Such regulatory intervention is necessary to protect consumers, especially those who may be more susceptible to certain unfair practices owing to reasons such as age, illiteracy, mental disability and poverty.
To ensure that regulators can understand and respond adequately to the different risks associated with innovation, they will need to step up their supervision and oversight capabilities. 208 This will require them to work with participants in the ecosystem to understand the products being offered as well as the attendant risks. This approach would ensure that regulators retain access to critical information in a fast-changing innovative marketplace, allowing them to adjust their regulatory policies where required. 209 In many cases, regulators will also need to work closely with legislative authorities to design appropriate legislative responses where necessary.

CONCLUSION
In developing countries with high mobile penetration rates, m-payments can assist in extending financial services to the unbanked populace. Consequently, the service has drawn the attention of regulators. While regulatory responses to innovation must be country-specific and tailored to suit local realities, policymakers seeking to leverage innovation for financial inclusion should learn from the experiences of other countries. Reflecting on Nigeria's experience, this article concludes that, while high mobile penetration rates are a strong indicator that m-payments may flourish in a given jurisdiction, their success largely depends on the regulatory environment.
More specifically, the article first argued that one reason for the dismal success of m-payments in Nigeria was the CBN's initial decision to exclude MNOs from providing m-payments services directly. Given their extensive agent networks, the article argued that the direct participation of MNOs in m-payments is crucial in driving efforts to improve inclusion. Recent regulatory changes support this view. Secondly, contrasting the CBN and CBK approaches to the early regulation of m-payments in their respective jurisdictions, the article highlighted the importance of flexible regulatory approaches that encourage communication between regulators and innovators. Such approaches may achieve a better balance between protecting the financial system and leveraging innovation to drive inclusion. Thirdly, while acknowledging policy shifts in Nigeria that allow MNOs to participate in the market through niche banking institutions, the article highlighted the difficulties associated with this approach by drawing on India's experience.
Fourthly, the article suggested that, if financial inclusion is a policy priority, regulators should embrace models that encourage investment by providers with good agent networks in underserved areas. Accordingly, it argued that MNO-led models or hybrid models are more likely to record transformative success. Finally, the article proposed that a pragmatic approach to regulating innovation is one that seeks to understand and identify associated risks to enable regulators to design proportionate responses.

CONFLICTS OF INTEREST
None 209 Zetzsche et al "Regulating a revolution", above at note 108 at 61-62.