Putting the Banking in Branchless Banking: Regulation and the Case for Interest-Bearing and Insured E-Money Savings Accounts

By Tilman Ehrbeck and Michael Tarazi, Consultative Group to Assist the Poor

Buoyed by a growing belief that poor people need a full array of financial services, financial inclusion advocates are now focusing on how to responsibly provide low-income individuals with financial services beyond microcredit—services such as savings, payments, and insurance. A primary obstacle to the provision of such services (particularly low-value payments and savings) has historically been high transaction costs. It has been too expensive to develop the infrastructure required to profitably reach underserved population segments. Branchless banking, however, is changing the economics of providing financial services by leveraging existing and widespread retail outlets and technology (particularly mobile telephones) to provide more services to more people at lower cost.

But when it comes to using branchless banking to provide savings services, regulation is often the key obstacle, particularly in the case of e-money issued by nonbanks (such as mobile network operators).[1] In an effort to distinguish such products from savings accounts, regulators around the world have regulated them as “payments” services, denying e-money accounts the benefit of interest payments and deposit insurance. In some cases, these prohibitions extend even to e-money issued by banks, particularly in countries such as the Philippines and Malaysia where e-money is regulated as a product regardless of whether the issuer is a bank or nonbank.

In regulating e-money as a payments product, regulators may be missing an opportunity to make great progress in financial inclusion. E-money can safely and efficiently be used as a savings vehicle, and regulators should therefore consider allowing e-money to offer the full benefit of savings accounts—interest anad deposit insurance—to the millions of low-income e-money users.

E-Money as Payments

E-money is commonly understood as (i) monetary value as represented by a claim on an issuer, (ii) stored on an electronic device, (iii) prepaid, (iv) accepted by third parties other than the issuer, and (v) redeemable in cash.

E-money can be issued by both banks and nonbanks. In the case of nonbank issuers, however, regulators typically safeguard the cash collected in exchange for electronic value (the “e-float”) by requiring 100 percent to be placed in an account held at a fully prudentially regulated bank.[2] This account is typically a pooled account that should be held in trust (or the equivalent) for the benefit of e-money customers so as to isolate the funds from claims by issuer creditors (in the case of issuer bankruptcy, for example), although other approaches also exist.[3]

Regulators have struggled with the business model. Collecting cash from the general public sounded like the equivalent of collecting deposits. However, in the banking laws of many countries, deposit-taking is considered the domain of fully prudentially regulated banks only. Consequently, regulators were in a bind. How could they permit nonbanks to collect deposits without requiring them to obtain banking licenses that would subject them to complicated and costly prudential requirements—requirements likely to prevent nonbanks from issuing e-money at all?

The solution was to regulate nonbank issued e-money as a “payments” product—focusing on the funds transfer functionalities of e-money and effectively lumping e-money issuers together with money transfer companies like Western Union. The cash-in function was not considered a deposit but simply the equivalent of handing money over to Western Union before its eventual transfer (within a prescribed time period) to another recipient. This “e-money as payments” approach was convenient not only for bankers struggling to avoid the question of deposit taking, but also for nonbank e-money issuers who had no desire to be licensed as fully prudentially regulated institutions and preferred to avoid unwanted attention from the banking sector as a result of appearing to compete on basic services.

While regulators should be commended for creatively enabling inclusive financial services, regulating e-money as a payment product may close the door to using e-money to provide the savings services that e-money more closely resembles. When an e-money customer gives cash in exchange for electronic value, there is typically no requirement that such electronic value ever be transferred.[4] In fact, many e-money customers use their electronic accounts as a means of safe storage. A 2007-2008 study of 350 M-PESA users revealed that for reasons of safety and convenient access, M-PESA is used as a storage mechanism by both the banked and unbanked (Morawczynski and Pickens 2009). [5] Another study found that M-PESA was used for both long- and short-term savings (Pulver 2008). It is this storage functionality that distinguishes e-money from a payments product and makes it more akin to a savings account than to a Western Union transfer.[6]

There is one significant difference however between e-money and a bank savings account. As long as the cash backing e-money is 100 percent held in a fully prudentially regulated institution, the e-money issuer does not intermediate the funds in a way that puts them at risk. [7] As a result, the e-float is not at any more material risk than cash held at a bank.[8]

E-Money as Savings

Once the intermediation risk of nonbank e-money issuance is removed, it is difficult to see why e-money could not provide low-income users the full benefits of a savings account. In fact, e-money accounts already provide one key benefit: safe storage. A secure mechanism for storing value is highly valued by users, particularly poor users with few safe options.

However, advocates of financial inclusion can do more than simply promote savings as safe storage. They can promote savings as interest-bearing and insured accounts—the type of savings enjoyed by most banking customers. In so doing, they can put the “banking” in branchless banking.

Distributing Interest. The payment of interest on e-money accounts provides several benefits to customers and regulators alike. For customers, interest encourages savings and teaches low-income users the time value of money. It affords many low-income users a rare opportunity to earn a return on their capital. Tellingly, when asked what additional service they would like to receive through M-PESA, users cited earning interest above all others (Pulver 2008, p. 5). [9]  For regulators, providing an added incentive to save encourages more citizens and more money to enter the formal and traceable economy, not only providing benefits to the financial system as a whole but also providing a means by which to monitor transactions in the fight against money laundering and terrorist financing.

Despite the benefits, no country currently permitting nonbanks to issue e-money allows such issuers to pay interest on e-float. Such prohibition often extends to “interest equivalents”—any benefit, such as free airtime, linked to the balance of the account. When pressed for a reason, regulators often simply state that paying interest is a banking activity. However, definitions of banking activity typically focus on taking deposits and, in most regulations, intermediating such deposits through on-lending. It is the intermediation of deposits that places the deposits at risk, thereby raising systemic concerns of the type prudential regulation is intended to mitigate. While nonbank e-money issuers are arguably taking deposits, these deposits (if 100 percent held in a bank) are not intermediated by the issuer. Even when regulation expressly cites the payment of interest as a banking activity, it is hard to precisely identify what risk lies in allowing nonbank issuers to pay interest.

There are several operational arguments against paying interest. For example, the interest accruing on such small amounts is often viewed as negligible. In fact, however, e-money structures present a unique opportunity for low-income individuals to earn higher amounts of interest. Because e-money accounts typically pool client funds for an extended period, the total balance often qualifies for higher interest rates than might otherwise be earned by low-value individual bank accounts. For example, a time deposit in Kenya currently earns 3.43 percent annual interest whereas, even if poor customers could open a bank savings account and meet minimum balance requirements, a savings account earns only 1.25 percent.[10]

Another argument against permitting nonbank issuers to pay interest is that it might cause such issuers to recklessly invest their working capital to provide greater interest rates to their customers. While this would not endanger the e-float kept in a custodial bank account secure from the issuer’s creditors, it could result in the bankruptcy of e-money issuers, ultimately putting the reputation of the entire e-money sector at risk. This argument presupposes that e-money issuers would compete on interest to the point of jeopardizing their entire business. Even if this possibility was not remote, one solution is to permit nonbank e-money issuers to simply “pass through” the interest already accruing on the e-float, rather than pay interest directly. This would benefit e-money users by encouraging issuers to negotiate with custodial banks for the highest interest rate—a benefit ultimately passed on to the customer.

An argument against an interest pass through is that e-money issuers need to keep the interest on the e-float for their own profit because they are not able to recoup costs through transaction fees alone. But this is an argument against mandating the payment of interest on e-money, not an argument against permitting it. Whether an issuer pays interest will ultimately be a business question based on whether the issuer can afford to pass through the interest (and assume the related administrative and technology costs). However, the success of M-PESA (in terms of number of customers and amounts transacted) suggests that the interest can be irrelevant to viability. Safaricom, the mobile network behind M-PESA, does not benefit from the interest accruing on deposited e-float. [11] (See Figure 1, M-PESA Legal Structure.) Instead, Safaricom and the Central Bank of Kenya agreed to donate the interest to charity rather than distribute the interest among the customers on whose funds the interest accrued.

Extending Deposit Insurance. Deposit insurance is meant to protect bank customers from a bank’s inability to pay its debts. Bank failures and the recent financial crisis have resulted in a rapid increase in the number of countries, currently 104 and rising, that have a government or private mechanism for insuring bank deposits.[12] These include a large number of poor and developing countries. Deposit insurance is not just for the benefit of bank customers. By encouraging trust in the formal banking system, governments promote savings, increase cash reserves, and stimulate the entire economy.

Yet, even though e-float is typically held in a bank,[13] few regulators in the developing world (if any) extend deposit insurance to customers of e-money accounts issued by nonbanks. The custodial accounts holding the e-float do benefit from deposit insurance. But because the funds are pooled, insured amounts are typically well below the e-float total. For example, the US$1,300 insurance limit in Kenya would do little to cover M-PESA’s e-float amount.[14] In addition, deposit insurance benefits the named holder of the account, which, in cases where e-float is not held in trust, is often the e-money issuer.

However, extending the benefit of deposit insurance to e-money is, in principle, a relatively simple endeavor—and the United States already provides such deposit protection. In the United States, as long as e-float is placed in an insured depository institution, it is considered a deposit and is insured. And the United States goes one step further: in the context of pooled custodial accounts, it affords pass-through protection to each individual customer up to the insurance limit. To qualify for pass-through protection, (i) the bank’s records must disclose the custodial nature of the pooled account, (ii) the records of the bank or the issuer must disclose the names of the individual owners and the amount owed to each owner, and (iii) the agreement between the issuer and the customers must indicate that ownership of the funds remains with the customer (Federal Deposit Insurance Corporation 2008).[15] These requirements are not difficult to meet and most e-money schemes already comply as part of their standard business practice.

Provided the pooled account is insured, pass-through deposit insurance need not increase the insurance premiums. Premiums are typically based on the total of domestic deposits, or in the case of the United States recently, the bank’s total asset base. Neither of these calculations would be affected by extending insurance protection to individual e-money holders since neither the total deposit nor asset base would change.


Banking regulators are understandably uncomfortable with nonbanks offering traditional banking services. The temptation is always there to insist on the centrality of banks, and models like Kenya’s M-KESHO take the pressure off of regulators to think about e-money’s potential role in promoting savings. A joint product of Safaricom and Equity Bank, M-KESHO provides M-PESA users with an interest bearing and insured Equity Bank account accessible through mobile phones. However, its value proposition for low-income customers has yet to be proven. For example, to withdraw funds from M-KESHO, a customer must first pay a fee to transfer funds from the M-KESHO account held at Equity Bank to the M-PESA account and then pay a second fee to withdraw cash from M-PESA. These two transaction fees largely undercut any interest benefit. M-KESHO is nevertheless promising inasmuch as it lays the rails for the cost-effective provision of other financial services such as credit and insurance. However the potential of models like M-KESHO does not obviate the need to explore how M-PESA and other e-money products may provide interest-bearing and insured savings more effectively on their own.

E-money represents a promising opportunity to provide low-income individuals with more than just payment and safe storage services: it can offer savings vehicles with the full benefit of interest and deposit insurance. And the extension of such benefits can be done with relative ease and at minimal risk. Nonbank-issued e-money products should not be seen as interlopers in the banking domain—but rather as a much needed stepping stone across which the benefits of high-quality savings instruments can be passed through to the millions who do not have access to them now.


Federal Deposit Insurance Corporation. 2008. “Stored Value Cards and Other Nontraditional Access Mechanisms, New General Counsel’s Opinion No. 8.” Financial Institution Letter. Washington, D.C.: FDIC, 13 November.

Morawczynski, Olga, and Mark Pickens. 2009. “Poor People Using Mobile Financial Services: Observations on Customer Usage and Impact from M-PESA.” Brief. Washington, D.C.: CGAP, August.

Pulver, Caroline. 2008. “The Performance and Impact of M-PESA: Preliminary Evidence from a Household Survey.” FSD Kenya.

Tarazi, Michael, and Paul Breloff. 2010. “Nonbank E-Money Issuers: Regulatory Approaches to Protecting Customer Funds.” Focus Note 63. Washington, D.C.: CGAP, July.

[1] Nonbanks are permitted to issue e-money in an increasing number of developed and developing nations, including the West African Union, Kenya, Rwanda, the Philippines, Malaysia, Indonesia, Fiji, and Cambodia. (And such an arrangement is under consideration in countries such as DRC and Burundi and in the South Pacific.)

[2] Note that this is a more stringent requirement than imposed on deposit-taking financial institutions, which are typically subject to reserve requirements mandating only some small portion of overall deposits to be kept in liquid form (typically cash) to satisfy potential depositor claims. This difference in treatment reflects a fundamental difference among banks, nonbank service providers, and their respective business models. A bank’s business is predicated on the ability to intermediate capital, i.e., take money from those who have it and provide it (in loans or other products) to those who need it. Nonbanks, on the other hand, are typically expressly prevented from intermediating deposits and thus must make money in other ways, such as transaction charges, lowered airtime distribution costs, and reduced customer churn.

[3] For a discussion on regulation of nonbank e-money issuers, see Tarazi and Breloff (2010). Some regulatory systems do not impose the requirement that the e-money float be held in trust, and some permit its investment in other safe and liquid investments, such as government obligations, rather than requiring it to be held in a bank. Still others are silent or ambiguous on the subject.

[4] Based on a review of applicable regulation and/or practice in Kenya, the Philippines, West Africa, Fiji, Afghanistan, and Malaysia. The authors did not conduct a global survey of applicable regulation.

[5] Nearly a third of banked customers and a fifth of unbanked customers use M-PESA to store value.

[6] In an effort to more firmly characterize e-money as a payments product, some regulators considered putting a time limit on the amount of time funds can be stored electronically. Perhaps realizing that such time limits could discourage use and savings, such time limits are not widely imposed, if at all. As e-money is used increasingly as savings, regulators may eventually feel compelled to impose such time limits.

[7] E-money issuers are often permitted to invest the float in government-issued securities—a form of intermediation considered lower risk. However, perhaps due to the lack of liquidity associated with such securities, most nonbank e-money issuers opt for the other legally prescribed option—holding the e-float in a fully prudentially regulated financial institution.

[8] The risk is further minimized in cases where the e-money issuer maintains the e-float in several banks, mitigating the risk of any one bank failing.

[9] The study showed that 38 percent of respondents cited earning interest and24 percent cited the ability to use M-PESA for ATM withdrawals (which has since been enabled).

[10] Interest rates as of January 2011. See www.centralbank.go.ke.

[11] Nevertheless, the terms of the trust do not expressly prevent Safaricom from benefitting from the interest. Section 6.2 of the Amended Trust Deed of 2008 states any interest “shall generally be applied first to defray the Trustee’s own costs…but may be applied for such other purposes (whether charitable or not) as the Trustee, may in its sole discretion determine.”

[12] International Association of Deposit Insurers at http://www.iadi.org/aboutiadi.aspx?id=79.

[13] Another safeguarding measure is private insurance. The European Union (EU), for example, permits safeguarding of funds backing e-money through insurance. EU Directive 2007/64/EC permits nonbank e-money issuers in lieu of liquidity provisions, to insure or comparably guarantee the funds backing e-float in an amount payable in the event that the nonbank issuer is unable to meet its financial obligations. EU Directive 2007/64/EC, Article 9.1(c) incorporated by reference from Article 7.1 of EU Directive 2009/110/EC (2009). It is not clear if any e-money issuer has availed itself of this private insurance option.

[14]In the case of M-PESA, the trust responsible for the e-float is charged 0.01 percent of deposited e-float as an insurance premium—though such expenses may be reimbursed out of interest accruing on the e-float.

[15] The second requirement that bank or issuer records disclose the names of the individual owners and amounts owed may require regulators to impose data storage and back-up protocols similar to those imposed on regulated financial institutions.

Click to comment


To Top