Blog

The Four Zones: A Missing Chapter in the Financial Inclusion Guidebook

Posted on by MSC

As of March 24, 2017, a little more than 9 million dormant PMJDY accounts were frozen, because the account holders had not transacted in over a year. A great deal of effort has gone into government initiatives to increase financial inclusion, but perforamance indicators such as dormancy in the PMJDY accounts is a reason for worry. The Government of India’s Economic Survey Report 2016-17 too finds that active use of PMJDY accounts is 40% (active defined as at least one transaction completed in the last 90 days). There could be many reasons for poor account activity – it could be that the product delivery is not quite right (the Economic Survey highlights that distance from nearest bank and the number of people per branch might explain poor account activity rates), or that the products themselves are not suitable.

We sought to answer this riddle with our usual customer-centric approach. We focused on the four personas that represent the majority of the mass market that has been the target for financial inclusion in India: a farmer, a daily wage labourer, a clerk who works in the city, and a security guard.

Shankar Singh is a daily wage earner. Most of his earning of about INR 8,000 (USD 117) per month is committed for household essentials and debt repayment. Since he is risk averse, and wants to avoid the hassles caused by distance, low literacy levels etc., he saves in a box at home. When the money runs out, Shankar is susceptible to falling into debt. Jai Prakash is a security guard who earns a salary of INR 8,000 (USD 117) per month and by the third week, when the money runs out, he turns to his employer for an advance on his salary and/or to his friends for interest free soft loans. He prefers the ease of such informal borrowing avenues and discounts the unreliable nature (in case the employer or friend is unable lend him money) or costs (e.g. interest paid) of these sources of cash. Sundar Devi is a dairy farmer and her daily income amounts to about INR 20,000 (USD 294) per month. In many months, this is enough to manage her household expenses and even put away some savings. Vijay Thakur, a salaried clerk, earns a little under INR 20,000 (USD 294) a month and meticulously saves, invests in insurance and takes loans for specific goals like child’s education, weddings etc. In a financially stressed situation, Sundar Devi and Vijay Thakur compromise on savings and investments, and may even fail to honour monthly loan repayment instalments. Such financial turbulence can be due to sudden spikes in unforeseen expenditure such as sudden illness, an investment in a business opportunity, a family function, or a life event, etc. (Read more about their lives here).

The two of the key determinants for all four personas’ financial decisions are the quantum and frequency of income – representing ‘how much’ and ‘how often’ for the mass market. These two together define the ability of a family to meet regular household expenses, planned future expense and/or any eventuality. These two factors also play a significant role in paving the way for four discreet financial phases through which a household oscillates.
A household’s disposable income influences the financial behaviour, which changes as the household moves through these four phases:

  1.   The comfort zone when money inflow is enough to meet household expenses;
  2.   The fluid zone when the household money has almost run out;
  3.   The stress zone when small loans are needed to smoothen consumption; and
  4.   The blast zone when credit is not enough, or simply unavailable, and liquidating household (or in the very extreme cases, business) assets is the only way to sustain basic household financial needs.

The mass-market keeps moving through these four zones and a closer look will provide a deeper understanding of the financial choices adopted by them to navigate through these phases.

The household in these zones might have a PMJDY account in a branch close to home, but still lack financial robustness and income security to use these accounts. The challenge, as we see here, is developing a versatile set of tools to service the mass market whilst accommodating the vari0us and changing money management strategies of different households. For instance, how can we offer Shankar a formal savings product that satisfies his liquidity requirements (accessibility), small loans for Jai Prakash without high interest rates, and better investment options for Sundar Devi and Vijay Thakur so that money in the boxes at home can be unlocked, nurtured and grown?

Expecting financial inclusion programmes to succeed by relying on easy credit, a bank account, or enabling hassle free payments oversimplifies the complexity of the challenge. None of these laudable initiatives has attempted to create the behavioural changes that are essential for financial inclusion. As a result, they have only seen limited success – either for specific customer segments or for a specific geography (for instance, there are financial services specifically targeted for migrant labourers, DBT beneficiaries and bill payers). The mass-market prefers tools that provide them with a tangible, trustworthy, and easily understandable approach to manage their money. They want clear visibility into their money, and (ideally) how it is working for them. So, for example, they see and know the whereabouts of creditor/debtor and friends/relatives. They are looking for tools that help them quickly replenish their liquidity when the need arises, ideally without trapping them in debt spirals.

While interacting with Shankar, Sundar Devi, Vijay, and Jai Prakash we kept asking ourselves if there is a product/service that can work for them both in the comfort zone and the stress zone. We did not find any. In fact, we found that existing financial services designs fall short of incorporating geographical, cultural, and literacy heterogeneity, critical to enhance usability. For example, leading mobile wallets rely on aesthetic abstract icon designs supported by text to serve its customers. However, when unassisted, such abstractions mean very little to an illiterate or a neo-literate person like Shankar and Jai Prakash. (Please see our recent work that identifies new dimensions to mobile wallet designs for illiterate and innumerate people).

We are at a juncture that calls for introspection on current design concepts to create a new breed of financial tools (not products) for the mass market. With myriad complex cognitive factors defining financial behaviour, creating products for specific customer segments or use cases will always have limited reach. Given the agenda of financial inclusion, service providers should create a suite of flexible and intuitive money management tools that cater to multiple use-cases, and meet the requirements in all the four zones for the multiple segments of mass market.

Understanding Karnataka’s Food Coupon System

Posted on by MSC

The Department of Food and Civil Supplies, Karnataka began reforming public distribution by rolling-out Food Coupons in Bangalore. While it was a well-meaning and thought-out idea it resulted in much heartburn among beneficiaries. While beneficiary distress was justified, the system itself was a large improvement over the one it replaced as it increased transparency and enabled better tracking of demand. The hiccup was only caused due to hurried implementation and poor communication – a common feature among many large scale government programs.

Learnings from Cash Economy for DFS Providers

Posted on by MSC

Shakuntala is a 48-year old housewife, staying with her husband, two sons, their wives and a grandson in Siswan village, Varanasi. Her husband manages all the household expenses. In his absence, she goes to buy groceries at times. She, however, is solely responsible for purchasing clothes and jewellery for all the family members during the marriage or any other festival. Any purchase she makes, Shakuntala needs to keep an account and has to inform her husband about the expenditure. Although Shakuntala has never attended school, she is adept in counting currencies and doing basic calculations using some very concrete ways – often used by oral people.

“Orality” refers to the modes of thinking, speaking and managing information in societies where technologies of literacy (especially writing and print) are unfamiliar to most people. Orality encompasses not just speech but a wide range of modes for personal and collective information management that are preferred to text in oral cultures – from pictures, tallies and cash, to apprenticeship, rituals and song.

There is a very large ‘oral’ population like Shakuntala (either illiterate or innumerate or both) who cannot rely on pen and paper-based calculations. This forces them to mental counting and reduces their speed, especially when dealing with large numbers, as they face difficulty in decoding numeric or arithmetic notation.

To better understand both the opportunities and limitations faced by oral people in using financial services, we conducted research in different villages across three states in India – Uttar Pradesh, Punjab and Bihar. We observed that oral people get maximum exposure to numbers when they start participating in cash economy (household or personal monetisation) since this is the time when they begin to apply basic concepts of mathematics, such as addition, subtraction, multiplication and division. This helps them in money management, business and household expense management, and largely governs their financial behaviour.

Our research findings help us strengthen this observation. We have considered three parameters to evaluate the numeracy capabilities of the oral segment:
1)    schooling,
2)    age, and
3)    gender

Basic mathematical problems based on real-life situations were asked to the oral respondents. In the first attempt, they had to calculate the problem mentally. In case the respondents gave incorrect answers or were unable to solve this problem, they were given a stack of cash to help them solve this problem. We constructed a simple indicator with a range of 0-7 (1 point for each right answer calculated mentally).

The average score was 4.2 for the sample. Neo-numerate respondents (5.0) scored substantially better than innumerate ones (3.9).

We found little correlation between mental numeracy skills and schooling. There were respondents like Shakuntala, who have never attended school but were prompt and accurate in answering questions related to basic addition, subtraction, multiplication and division. These respondents however faced problems when our team asked questions testing complex calculations, including interest rates1 and  division.2  

Older respondents (especially in the age bracket of 46-60 years) outscored younger ones. The question on long division was a bit complex in nature and involved some application of estimation. A direct relationship was observed among the percentage of people answering it correctly and their age-group, i.e. a high percentage of people in the higher age-group answered the question correctly.

On average, young people in the sample had weaker mental calculating skills than unschooled people, and the 16 respondents aged over 45 had mental calculating skills that were equal to those of the 19 respondents who had completed one or more years of post-primary schooling.

Females scored better than expected (3.9 out of 7) as most of them managed their household expenses and were exposed to conducting basic calculations.

What does it tell us?

Individuals who are more actively involved in the economy (either earning money outside their home or actively managing household expenses) have less trouble dealing with numbers. Continual practice sustains or can even enhance their cognitive capabilities. For example, a labourer who paints houses must calculate input costs per square feet. A vegetable vendor must calculate price-per-quantity of different vegetables bought and sold (including spoilage) and all-in profit margin. A housewife who handles household expenses must buy groceries for the family, pay electricity bills and calculate/track total household income and expenditure. Thus, through the frequent necessity of processing numbers, costs and cash, people become quick in calculations.

Like Shakuntala, Ramesh is an oral person. He was asked a question on subtraction: You are in the market on Saturday morning, and you wish to buy a bag of rice from the shop that costs ₹780. How much change should you get from ₹1,000? 

He did a mental calculation and arrived at the answer ₹320. Since he was unable to give the correct answer at the first attempt, he was handed ₹5,025 and asked to recalculate. Within a minute he did the calculation with the help of the currency and answered ₹220, correcting himself. When asked how he calculated the first time, Ramesh explained that while calculating mentally he subtracted using two steps: (i) ₹1000 from ₹700 and (ii) ₹100 from ₹80, thus arriving at ₹300 and ₹20 respectively and then he added these two numbers. He was not able to comprehend that ₹780 was a number closer to ₹800 and hence the approximate answer should have been slightly larger than ₹200.

In our field work, we found that cash was being used by oral adults as a kind of simple calculator, which gave them access to ways to calculate larger numbers than they could otherwise reach. The paper rupee banknote offers oral individuals many clues to skills that support microenterprise and household management. For example, cash involves:3

  • standardised nominal values, with standardised relationships between them, which refer and evoke the Indo-Arabic number system at the base of the financial system;
  • an explicit formal zero and place value;
  • denomination values that operate in the useful reference range of the local economy and facilitate rapid large-number counting and processing; and
  • a highly developed and established ecosystem of institutions and cultural practices and habits.

This means that the cash economy has become an informal school and reference points for mental numeracy. In this ‘school’, the incentives to pay attention in class are very high.

Does DFS solution provide learning for the oral segment as cash does? 

In India, as in many countries, the cash economy appears to be more accessible to the oral segment than the financial system and economically active individuals respond by acquiring surprisingly advanced mental numeracy skills. However, their inability to read arithmetic notation restricts them from directly accessing financial services as they may still have to depend on their literate counterparts. With the advent of technology and mobile wallets being introduced, people today have access to easy and convenient banking services. Since these wallets are highly text-intensive and filled with abstract icons, their usage is limited to the literate segment and does not extend to the oral segment.

An understanding into orality in India does provide a great understanding of why some of the digital innovations (digital wallets, payment interfaces) did not achieve mass scale adoption. In the same line, it triggers the thought that digital interfaces need to be carefully designed, considering the behaviour of oral people.

 1Question on interest rate: What is 100% of INR 1,000?

2Question on  division: If you have to set a goal of saving ₹50,000 in 5 years. How much do you have to save each month to reach your goal?

 3Matthews, Brett Hudson (2016). Oral Financial Numeracy, My Oral Village Inc., Toronto, p. 26.

Lessons from Orality for Digital Financial Services Development

Posted on by MSC

At 6:30 a.m.,in a busy wholesale vegetable market on the outskirts of Varanasi, Shanti Devi is haggling with the wholesale vendor of cauliflower. The price is negotiated at ₹300 a sack. Shanti takes out two ₹500 currency notes from her batua (cloth purse) and hands it to the vendor for two sacks. The vendor hands her ₹400 in change, which she quickly confirms mentally.

Shanti has not been to school and is not able to decode large written numbers (3 or more digits in length) or text. The numbers and text written on ₹1,000 and ₹500 notes do not communicate their value to her.

How does an oral businessperson like Shanti Devi calculate the price of cauliflower, count currency, or even recognise currency when she hardly knows the numbers or alphabet? Probing into these will lead us to a language that is understood by not only her but also millions of people like her.I

We asked a sample of oral people like Shanti Devi to perform certain tasks, like recognising currencies, counting money and performing some arithmetic calculations. Our observations offer useful insights into how this segment should be treated by financial services providers (FSPs).

Recognising currency denominations

Shanti Devi could not read written numbers but she could easily recognise any currency given to her. She cross-checks colour, relative note size and imagery against each other. She may also observe the number of zeros in the number printed on the note. Taken together, these clues are always adequate for Shanti to total the amount, though individually they are not.

  • When fake currency notes were given to Shanti and other respondents with the amount clearly marked but without correct colours or relative sizes, participants mistook the ₹1,000 for the ₹20, as the colour of both the fake notes was similar.II
  • Asked to count fake currency notes of different denominations but the same size, some confused the two denominations with each other.

Reliance on redundant clues is common among school learners of arithmetic. FSPs can provide information on large numbers in multiple formats: for example, through voice on a mobile phone, or images of currency notes accurately and proportionately rendered, to enhance confidence.

Counting money 

We asked 56 individuals like Shanti Devi to count ₹5,025, in a stash of 40 notes and coins. Based on initial screenings by our team, not a single person could decode a 4-digit numeral string (like ‘4,702’ or ‘5,097’). Nevertheless, 35 respondents (63%) accurately counted the cash. However, they varied widely in skill – some took longer and were more likely to err, while others counted quickly and rarely made a mistake.

  • The least experienced tried to keep running cash totals, without organising the notes first. This method is stressful and rarely yields a correct answer.
  • More experienced counters have learned to organise notes in groups, which they can then add together. Nevertheless, to reach a correct total, these groups of notes also have to be easily processed: the sub-totals have to be few and easy to remember.
  • Skilled counters first form one high-value group with an easy total to remember (such as ₹4,000), and then clean up the rest when this essential sub-total is securely memorised.
  • Skilled counters try to chunk currency in 10’s, 100’s or 1,000’s as addition or subtraction of one is the easiest.

For oral individuals, cash is a bridge to large-number understanding. Those who cannot read Indo-Arabic code (e.g. ‘4,702’) can arrive at the number using cash notes and coins. However, the speed at which they do this varies considerably. FSPs should not expect oral customers to process large numbers quickly. If given time, however, they often succeed, increasing their confidence.

Calculations

Shanti bought two sacks full of cauliflower and each sack had around 48 to 52 cauliflowers. The total cost of the two sacks is ₹600. She knew that to make a profit she must sell the cauliflowers at ₹10 apiece. She also knew that some pieces would sell at ₹12 to ₹15 while some will sell at ₹8 to ₹10 apiece. After some waste, she expects her total revenue to be around ₹1,000 and profit of around ₹200 a sack.

Respondents were surprisingly good with these sorts of calculations. We asked them three questions. The first question involved a mental calculation of 5 X 25. The second involved division of 50,000 by 5, while the third involved division of 50,000 by 60. Most were able to answer the first two questions accurately, and nearly half came within a reasonable estimate range of answering the third (between 750 and 900).

When cash is at stake, errors are very stressful, so there are significant incentives to learn. Neuroscientists have found that the human mind has very limited ‘working memory’.III  In school, we learn facts like the times-tables and algorithms like multiplication and division by constant repetition and practice. These mathematical tools economise on our working memory, offering very powerful shortcuts that support fast, accurate and low-stress numeric processing.

Without these tools, unschooled adults use cash as a simple calculator  that gives them access to ways to calculate larger numbers than they could otherwise reach. Without multiplication-tables, for example, working out the value of seven ₹500 notes cannot be done by multiplying “5X7”. Instead, one will count “500+500+500+500+500+500+500”. This process is longer and involves greater risk of error (“how many “500’s” have I added so far?”), especially under time pressure. When available, rupee notes will be used to support calculations.

Lessons for FSPs

While developing digital financial products and services for the oral market segment, FSPs can adapt their interfaces in ways that accommodate oral habits and practices. Oral customers may use multiple clues to decode large numbers, and with time, they may graduate to more advanced cognitive tools provided through schooling.

In fact, such an approach can be used for all facets of financial services development. For example, while designing financial literacy campaigns, designers can build on ways that people like Shanti manage money. See also our pitchbook on ‘MoWo mobile wallet design for the oral segment’. 

 

 iAs per latest data, around 264 million adults, aged 15 or more, in India are illiterate.i There are also many millions of ‘neo-literate’ individuals with very weak reading and writing skills. Illiterate and neo-literate individuals together form the ‘oral’ market segment, of which nearly two thirds are women. “Orality” refers to the modes of thinking, speaking and managing information in societies where technologies of literacy (especially writing and print) are unfamiliar to most people. Orality encompasses not just speech but a wide range of modes for personal and collective information management that are preferred to text in oral cultures – from pictures, tallies and cash, to apprenticeship, rituals and song.
iiFieldwork was done in pre-demonetisation period.
iiiThe classic paper on working memory is Baddeley, Allan D., and Hitch, J.G. (1974). Working memory. In G.H. Bower (ed). The Psychology of Learning and Motivation: Advances in Research and Theory. Vol. 8, pp. 47-89. Academic Press, New York.

Digital wallet adoption for the oral segment in India: Concept development for MoWo (Mobile Wallet for Oral)

Posted on by MSC
A lot of financial inclusion efforts around the globe are targeted towards people who are new to technology and perhaps new to the formal financial system. Many of these people belong to Oral segment of society. “Oral” people are not comfortable with written numbers, they also do not know how to read and rely on mental calculation when it comes to any mathematics. Learning about them is critical for all of us who are working to make digital financial services work for financial inclusion, as these people form the financial excluded segment largely.
The objective of this research was to develop conceptual wireframe of a mobile wallet for ‘oral’ people to use. The user interface was developed keeping in mind the cognitive usability constraints of oral segment.
The result of our work is first ever wallet design for Oral ; that’s what we call it –  MoWO – mobile wallet for oral. We believe MoWO is the first step in the right direction when it comes to financial inclusion of Oral people.

Demand and Supply Side Challenges and Potential Opportunities for Agency Banking in Uganda

Posted on by MSC

The Financial Institutions (Amendment) Act 2016, enabled agency banking in Uganda. What can financial institutions launching agency banking learn from existing data on the banking sector, from experience of mobile money, and from client expectations. The Financial Sector Deepening Programme (FSD) Uganda, commissioned MicroSave to find out.

Teams from MicroSave conducted interviews and focus group discussions with bankers, the banked and unbanked, mobile money users, mobile money agents and potential bank agents, mobile network operators and the Uganda Bankers Association.  The team studied data available on the Ugandan financial sector including the FinScope Survey 2013, the Intermedia Survey 2015, MicroSave’s Agent Network Accelerator Survey 2015 (from the Helix Institute). Respondents were from rural and urban areas in four research areas.

The case for agency banking: The FinScope 2013 survey suggests that 58% of Uganda’s adult population are potential users of mobile money; of this there are over 14 million unbanked adults. But the startling statistic is that more than 9.3 million adult Ugandans at the time of the survey lived more than one hour away from their nearest bank branch; most of these, some 8.6 million live in rural areas. Well implemented, agency banking can bring a range of banking services to Ugandans who have never had a realistic option of a bank account. Conversely for financial institutions, agency banking offers the potential to onboard, and provide services to large numbers of new customers.

Reaching out to large numbers of customers is feasible. The FinScope (2013) survey shows most transactions occurring in the banking hall are those which can be performed by agents. The leading transactions are over the counter withdrawals (13.8% of total transactions), ATM withdrawals (8.8%) and cash deposits (18%), so called cash in – cash out (CICO) transactions. Information requests (5.9%), receiving money (4.4%) make up the most frequent transactions. This transactional banking can all move to agency banking.

Ensuring a Customer Value Proposition: Focus Group Discussions with bank customers showed, why customers chose existing banking channels – the banking hall was preferred for larger amounts, it was a trusted, safe environment, it had constant liquidity. The ATM was fast, convenient and less costly, mobile banking was easily accessible. Adoption of agency banking should then focus on convenience, cost, a safe environment and liquidity.

Care needs to be taken to price transactions appropriately, banked customers were concerned that some mobile money transactions involving bank to wallet transactions, followed by cashing out were very expensive. Institutions will need to find a pricing point that appeals to customers, and still provides sufficient incentive for agents.

Customers valued bank branches as points to resolve customer care issues, as a place to access information on products and services and a point to access loans; but they felt that lengthy queues, high charges, and system downtime were issues which needed to be resolved. These observations are more problematic for agency banking. Financial institutions will need to invest heavily to provide consistent uptime, expand call centre operations, staff to support reversals and customer onboarding, and train agents to provide quality services. Banks may find it more difficult to sell non transaction based services through agents, however, pulling transactions from banking halls should make it easier for financial institutions to sell more complex products through their own outlets.

Service Offerings: Research with customers and financial institutions suggested that the initial service offerings customers will value the most will be cash in, cash out, facilitating account opening, and credit. Each of these services places specific demands on financial institutions. Managing cash in and cash out requires strategies to monitor and manage cash and float at agents alongside careful agent selection and the creation of appropriate mechanisms to rebalance cash such as intermediaries or so called super agents. According to draft regulations, bank agents cannot open accounts – responsibility for this belongs with the bank, but they can facilitate account opening. However, experience from other countries shows that this works well where there is a National Identification Card where there is a central database of client names and addresses which can be pulled digitally upon request. This is possible in neighbouring Kenya, free of charge for banks, a feature which has helped to grow financial access by making onboarding new customers as easy, simple and cost effective as possible.

The final service requested by potential customers, credit, requires unpacking. Once again agents are not positioned to offer credit services on behalf of banks, but they can, for smaller loan amounts disburse cash, they could be a collection point for information or initial applications and a repayment point for loans (through cash in transactions). The innovation that for many customers would link credit to agents, is mobile phone based instant loans, like M-PESA’s MShwari or MTNs Mocash. Through data analytics financial institutions develop algorithms through which they can provide credit lines to customers, which in turn the customers can access through bank agents.

 

Customer Preferences for Agents: Potential users surveyed showed a distinct preference for certain types of agents, preferences included petrol stations, SACCOs and grocery shops. However, financial institutions will need to carefully consider that decisions on petrol station involvement are often taken by the franchise owners (rather than the franchisees), and that access to cash in petrol stations is carefully controlled through drop boxes which staff cannot access.  SACCO performance in Uganda varies considerably, and screening criteria will need to be developed to source the best SACCOs. Grocery shops are ubiquitous, but tend to struggle more with liquidity management and rebalancing.

Potential users choices for which agents to use were being driven by trust (security), liquidity, opening hours, convenience and space; the factors driving SACCO inclusion were considered to be liquidity and convenience. Potential user preferences for agents included the agent being fluent in the local language plus English, and to have safe, secure, permanent premises. Focus group discussions indicated a preference for SMS receipts amongst non-banked and a preference for a physical receipt amongst the banked.

Factors Promoting Agency Banking: The factors promoting agency banking differ according to the group – banked users appreciate the convenience likely to be experienced by agency banking, whilst non-banked users are looking for less costly services. Banked users noted reductions in risk from carrying cash long distances, and reducing the cost of transport to the bank.

Will Agency Banking Increase Movement Between Institutions? The sample size is too small to conclusively answer the question of whether the introduction of agency banking will increase movement between financial institutions. Research suggested the factors which might keep people with an institution included the challenges of opening another account, the fear of untested services, and the existing history (and trust) that had been built with an institution. However, agency banking, combined with easier onboarding due to the National Identity Card offers the potential for customers to easily switch between institutions or to hold multiple accounts. The financial education which may accompany the launch of agency banking, combined with peer to peer learning, may further decrease friction and encourage competition between financial institutions.

The Concerns of Potential Customers: Potential customers were most concerned about fraud, and illiquidity, followed by network instability and the lack of electricity. These concerns have been noted in other survey’s such as the Intermedia Survey (2015) and MicroSave’s Agent Network Accelerator Survey (2016) from the Helix Institute. In fact Uganda has the highest incidence of mobile money fraud reported by agents out of ten countries in which MicroSave has conducted extensive surveys. This points to the importance for Uganda’s financial institutions in developing risk management frameworks, and risk mitigation strategies. The very fact that banks are often perceived to be safer than mobile network operators may offer banks an advantage in agency banking over their mobile money counterparts.

Likely Characteristics of Successful Bank Agents: The Agent Network Accelerator Survey indicates whilst most agents have less than a year’s experience in mobile money, the most successful mobile money agents are more established in their businesses. The research suggested that mobile money agents were less likely to have audited books of account and to pay tax on their business than more mature businesses; factors which will need to be considered during agent selection.

Understanding the Expectations of Agents: The biggest motivators for potential agents were commission income, the potential to cross sell other services to agency banking customers, and to create employment. At the same time, potential agents have expectations of banks, the most frequent expectation was for training, the second was for commission income, the third for sensitisation of end users, and to provide the necessary tools of the trade. Other slightly less requested asks included supervision, working capital loans, and liquidity management support. Financial institutions introducing agency banking will need to develop a package to meet these expectations. Most mobile money agents felt that banks would have tighter compliance requirements than mobile money operators, but felt that they would be able to meet them.

Factors for Success from an Agent Perspective: Whether mobile money agents, or established businesses, the most common success factor given by respondents was a high public awareness campaign, the second a good network signal. The established mobile money agents showed their perceptions of fraud, by also listing currency detectors, and video cameras.

Motivations for Banks: The most common motivations for banks considering agency banking is client acquisition and increasing activity on customer accounts. Additional motivations included decreasing cost to income ratios through better utilisation of bank assets, and to decrease congestion in banking halls, and finally to decrease the cost of deposits. All banks believe that numbers of accounts will increase, some think this will be a rapid change, others more gradual, the majority of the respondent banks feel that dormancy will decrease, and that new product development focusing on the channel will be important. The anticipation is that the launch offerings will be similar, including cash in cash out, balance enquiries, bill payment, facilitating account opening, loan repayments and loan disbursements. Thereafter innovations around credit are anticipated.

Enabling Customer Recourse: Agency banking by nature utilises networks of third parties whether agents, merchants, system vendors, super-agents, aggregators, and other financial institutions. This network of relationships and reliance upon third parties creates many points at which transactions can fail; in addition to the mistakes made by customers, and deliberate fraud. These circumstances call for advanced customer service recourse. Sector respondents identified a range of options for customer recourse, including agency banking supervisors in branches, call centres – with numbers displayed in merchants. However, respondents identified empowering agents, enabling them to be a first point of contact as an important resource, to do this, careful selection and training of agents is required, alongside supervision and refresher training. Toll free call centres are also important to facilitate recourse for all customers