Time for Cash to Cash Out? Wall Street Journal 11 February 2012 Scoping Kenya s Path to a Cash Lite Society July 2012
1. It s costly. What s wrong with cash? An electronic payment costs 1/3-1/2 of a paper payment. Estimates from US, Australia, and EU project 1-2% GDP in cost savings from shift to electronic payments. 2% of Kenya s GDP (2010) is about Kshs 29 billion. European Payments Council Humphrey, D; Willesson, M; Lindblom, T; and Bergendahl, G (2003)
What s wrong with cash? Transaction costs at a range of SSA financial service providers Bank A (Rural) Bank B (Urban) Bank C (Rural & Urban) M-PESA (Rural & Urban) Time in banking hall (minutes) 32 87 21 8 Time to travel (minutes) 17 19 22 11 Cost of travel (one way) $0.41 $0.20 $0.27 $0.04 % of income over a month 7% 2% 4.5% 1.3%
And are no help in making financial service business models work in the low-income segment. $3.00 $2.00 $1.00 $0.00 Monthly net income per transactional account by balance quartile (US$) $2.18 -$1.00 -$2.00 -$3.00 -$1.30 -$1.62 -$2.06 Quartile 1 Quartile 2 Quartile 3 Quartile 4 Source: Stylized product model based on actual portfolios of underlying transactional savings accounts at 8 banks; BFA 2011 *Avg. bal: Q1 = US$ 1, Q2 = US$ 10.98, Q3 = US$ 33.28 Q4 = US$ 1331 5
What s wrong with cash? 2. It s vulnerable. Causes losses/theft within business handling Consumer theft, loss, damage in fire, flood, etc. Prone to graft-- untraceable Possibility of counterfeit In the USA, an estimated 14% of income is hidden from tax authorities due to unreported cash income. What might that level be in Kenya?
What s wrong with cash? 3. It s restricting. Required geographic proximate transactions where travel may be necessary to transact and where prices or services may be unfavorable When ING started offering online-only banking thanks to pervasive e-money in the US all people had access to a new, attractive savings option regardless of their physical address.
8 Imagine a more financially connected world Social and economic development is largely about connecting people to ideas and opportunities around them why shouldn t we apply that same platform logic to financial services? Mas & Porteous (2012) In 1905, a third of US households had access to grid electricity. A century later, close to 80% of the world s population is connected to the electrical grid.
9 Living in an i-fi world i i nclusive i nterconnected i nformation Fi Fidelity=trusted It is an inclusive cash lite world. NOT cashless but 1. Every person has an electronic account with a regulated institution 2. Electronic transfers are real time and close to free 3. Application/ content providers thrive on information rich products
Does anyone yet live in an i-fi world? Average US consumer share of payments per month, 2009 30.7 US consumer now uses cash and cheques less frequently than electronic payments. 6.2 18.4 8.7 UK (2010) Median cash transaction $5 94% have a plastic payment card; debit card holder used 15X per month vs cash 32X. Gradual decrease in volume of cash payments but it is still double the number of all e- payments. 20% relied exclusively on cash. Cash Payment card Cheques E-payments Source: Boston Fed Survey of Consumer Payment Choice (2011) Source: The Future for Cash in The UK, Payment Council (2010) 10
11 How to get there (an i-fi world): There are at least three necessary components 1. Infrastructure: Build bridges to cash 2. Instruments: Make e-money useful to consumers 3. Products: Connect with consumer mental models
FSD-Sponsored Study: Purpose: Understand cash movement in Kenya to identify research needed to guide a transition to a cashlite economy in which technological solutions dramatically reduce costs of cash, and those benefits reach the poorest and most vulnerable segments of Kenyan society. Research Questions 1. How does money (cash and e-money) move through a small, fairly contained economy? 2. What are the current payment habits of different strata of this community s economy: producers, consumers, service providers, merchants, and financial institutions and aggregators? 12 12
Two Current Views of the Kenyan Payment Space Consumer Perspective More hopeful for i-fi G2P P2P G2B Integrated Payments Space B2G Payment Silos B2B B2P P2G P2B The nature of this payments universe is determined by: Consumer perceptions Product features De facto integration
Core Site, Linked Market, and County Hub 1.25 hours, KES 150 one way for transport Core Site About 8,000 people; No banks or chains Most households engaged in some form of agriculture County Hub About 200,000 people; Paved road connects to NBO Home to bank branches, major distributors, and some national chains (ex: Naivas) 45 minutes, KES 100 one way for transport 30 minutes, KES 90 one way for transport Linked Market About 4,000 people; No banks or chains Major livestock market
15 Two paths towards cash-lite : 1. Increase local recirculation of cash (reduce cost of moving large amounts of cash over space) Bridges to Cash 2. Increase prevalence of e- payments to decrease use of cash in total. E-Payments M-PESA & Bank Agents $ E-value Already happening on pretty significant scale: mobile merchants deposit at M- PESA after market-day sales. Cannot eliminate costs of transporting large sums of cash, because 1) Agents are incapable of large transactions for security reasons 2) We cannot assume cash equilibrium. Cashless (but not necessarily electronic) payments already account for 65% by value of non-local stock (purchases for resale) purchases. But, VERY uncommon for most consumer transactions. A single mango trader may pay out KES 400,000 in a day to local producers in cash. No local M- PESA or bank agent holds more than KES 200,000/day or accepts withdrawals >KES 50,000.
Bridges to Cash All about recirculating money locally
CORE SITE: Yes, there is a lot of movement, but quite a lot of cash stays local. 17 About 85-90% of cash transactions are local for poor households* About 60% local for richer households* Airtime Labor (farm and business) Food (especially among poor) Clothing (especially among poor) Rent for homes and businesses Stock for smaller traders Transportation (large part) Water Chamas
Overview of cash (not value) movements out to linked geographic zones 18 Cash out Smaller locations/ sub-locations surrounding site Stock purchases by visit and mobile distributors Mobile vendors Health care Shopping by salaried people/wealthy Asset purchases Banking Fuel Entertainment Core Site Linked Market Stock purchases (small value in cash) County Hub More distant grain producers Mombasa/ Nairobi Cereal purchases in times of shortage
*Extrapolated from merchant census quantitative data during a period when site was a net buyer of cereals. 19 CORE SITE: Summary of cash leakage channels (ranked by magnitude) Cereal purchases in times of shortage Stock purchases by visiting and mobile distributors Mobile market vendors Stock purchases by merchants purchasing direct from outer towns and cities Health care Shopping by salaried people/wealthy Asset purchases Banking Fuel Account for KES 6.6 million leaving core site in cash every month*
Most large value of non-local stock purchases is already done cashlessly. That KES 6.6 million in cash leakage from stock purchases is only 35% of the value of non-local stock purchases. Variable Sum (Entire Site) Mean (per merchant) Median (per merchant) Total monthly spending on stock 21,900,000 116,345 12,000 Total monthly value of supply purchased in cash 9,589,940 51,010 9,000 Total stock purchased in cash going outside core site 6,549,180 34,836 0 Total value of non local stock purchases 18,800,000 99,969 1,000 Local suppliers account for 33% of all suppliers to the core site, but only 14% of the value. Larger values of stock (for mostly larger businesses) are purchased from outside the core site. 20 *Analyzed from merchant census quantitative data. 20
Most stock purchases by number are done in cash, but these tend to be lower value. Mobile money and cheques tend to be used for larger purchases. 100% 80% 60% 40% 20% 0% 90% KES 2,000 KES 7,000 6% KES 20,000 KES 2,000 2% 3% Cash Mobile money Credit Bank cheque % Suppliers paid this way Median tx size KES 25,000 KES 20,000 KES 15,000 KES 10,000 KES 5,000 KES 0 Median tx size *Analyzed from merchant census quantitative data. 21
Overview of cash (not value) movements in from linked geographic zones 22 Cash in Saturday market day shopping Smaller locations/ sub-locations surrounding site Core Site Saturday market day shopping Cattle sales Linked Market MPESA daily brings cash Teachers/ salaried workers bring portion of salary in cash from bank/sacco Schools bring cash from bank County Hub Mango buyers Cereal traders during harvest Portion of remittances carried in cash (mostly Mombasa) Medical visitors Mombasa/ Nairobi More distant grain producers
23 MPESA agents and agricultural product buyers appear to be primary suppliers of large cash inflows. Each of 13 MPESA agents brings about KES440,000 in cash per week to accommodate withdrawals* KES 22.9 million shillings per month Often transported via motorbike or matatu with no security; very limited concerns re: security of cash Grain and mango traders pay each farming household about 20,000-100,000 per season in cash in one payment**
24 Estimated net inflows can and do appear to vary significantly over the course of the year. Making some estimations on actual aggregate flow values, we see a trend something like this: Businesses have enormous seasonal fluctuations, and so do remittances. Urban senders report sending more frequently during the dry/hungry season. 30,000,000 25,000,000 20,000,000 15,000,000 10,000,000 5,000,000 0-5,000,000-10,000,000-15,000,000-20,000,000 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Cash In Cash Out Net Cash Inflow
25 Bank and M-PESA agents help keep cash local, but they cannot accommodate large deposits and withdrawals, which are Heavy /expensive for the economy. 200,000 Maximum Transaction Sizes for Cash in/cash out 200,000 150,000 100,000 50,000 30,000 50,000 70,000 75,000 100,000 0 Maximum Deposit avg- =5,000 Maximum Withdrawal avg- =2,300 Maximum Cash Held M-PESA Agents Equity Agent
Electronic Payments Converting transactions to more efficient electronic/mobile mechanisms
Conceptual overview of cash movement between tiers 27 Central Bank of Kenya Financial Institutions Informal Formal Suppliers Merchants/Service Providers Consumers
Consumers How people are paid The only substantial non-cash income payments are for formal workers and remittances. Agriculture payments stand out, because they are large, infrequent, and yet done in cash. Type of Income Median Pmt Frequency Median value % in Cash % MPESA % Bank water sale 333 225 100% 0% 0% boda boda 300 325 100% 0% 0% business 48 1,500 92% 8% 0% informal work 25 300 100% 0% 0% plowing 15 2,000 100% 0% 0% remittance 12 2,000 16% 79% 5% teaching 12 20,000 0% 0% 100% agriculture 2 5,000* 100% 0% 0% dividend 1 14,000 0% 0% 100% *There is a wide range here from households that sold one bag of maize last year and those who sold several or sold mangoes and have large, lumpy revenues. 28
Consumers How people are paid And people PREFER to be paid in fairly lumpy values electronically than in cash or in small values. Daily workers would prefer to be paid monthly. They say that small value payments make it too easy to spend they don t value the payment as much and feel it s easier to be wasted. Respondents told us that electronic payments institute discipline, because the act of having to go withdraw forces you to think about what you need and have at least a rough budget in mind. 29
The Foundation: Consumers How consumers pay* The majority of transactions are very small, under KES100 Lumpy payments (>KES500) tend to happen in the following situations: Formal salaried households shopping in large supermarkets once monthly in county hub Secondary school fees, especially when paid by wealthier households (poor families pay in small increments) Remittances sent (already electronic) Business license payments Some medical expenses (though often in small increments) 30
Frequency of transaction per year 1200 1000 There is a clear tradeoff between transaction frequency and size. 800 600 400 200 0 Food Airtime Water Church Appx. frequency # per year Posho mill Hospital Transportation Clinic Haircuts Fuel Primary school fees Secondary school fees Laborers Appx TX Size Remittances Clothing Agriculture inputs Goat purchases Licenses 4500 4000 3500 3000 2500 2000 1500 1000 500 0 Approximate transaction size (KES) 31
Consumers: Rich vs. Poor Cash handling differs significantly across richer/formally employed and poorer/informal households. Inflow Richer/More Formal Earners Larger, lumpy, predictable income. Receives salary into financial institution in county hub. Leverage salary for large, infrequent loans from formal institutions. Poorer, Informal Earners Frequent, small cash inflows for work. Receive larger, less frequent income from remittances paid via M-PESA, but cash withdrawn immediately and in full. Agricultural produce sales done once per year in 1-3 transactions and paid cash by single trader. Outflow Spend much greater share of earnings in town Larger transaction sizes Less frequent transactions for day-today needs, like cereals Pay school fees in bulk using cashless method (pay per term) *20% HHs interviewed earned <KES 10,000/mo; 60% at <KES 30,000 Frequent, very small transactions for day to day purchases. Larger payments (like clothing and school fees) divided into small cash installments. School fees paid daily or weekly. Because of small, daily expenses, reluctant to lock up funds in a financial institution. 32
33 Our census identified 188 Merchants. Other 26% Duka (small) 16% 4% Hawker (at dedicated site) 5% Spare Parts Store 3% Cafeteria 3% Butchery Clothing Shop 3% Mali Mali 4% Green grocer 16% Open Market Stall 14% Boutique 6%
MERCHANTS: Median merchant monthly revenue is around KES22,000. Larger merchants source more of their stock from outside the Core Site. Variable Sum Entire site Mean (per merchant) Median (per merchant) Avg. estimated monthly revenue for all merchants 12,000,000 63,786 22,282 % of shops with 100% local suppliers 36% % of shops with at least one non-local supplier 64% Avg. estimated monthly revenue for locally supplied merchants 2,515,016 37,538 10,440 Avg. estimated monthly revenue for merchants with non-local suppliers 9,476,900 78,322 27,380 Merchants with highest revenue sell hardware and cereal. They also tend to make large purchases of stock for resale with non-local suppliers. Data from merchant census. 34
35 Saturday market traders have some unique features. Vendors come from all over the county to sell their products here. Most visit multiple markets every week, hitting key markets throughout the district or county over the course of the week. Median daily cash revenues on average day=ksh3000 Important source of cash outflow unless they deposit locally.
36 While about ½ of market traders leave with cash, the other half instead actually does deposit cash locally at a bank or M-PESA agent. Where do you deposit your business cash? Non-Market Merchants Market Merchants Bank branch 19% 9% Bank agent 9% 6% MPESA* 24% 44% House 45% 41% But, since only small scale traders, only blocks an estimated KES 56,096 from leaving the community. *Across both merchants and consumers, we were told that funds deposited on M-PESA were for very short term cash flow needs and small emergencies. Banks are used only to save towards a goal. Data from merchant census.
When it comes to retail payments, we see very similar patterns, even when we look rigorously at a larger sample of merchants. Gates Foundation-Sponsored Retail Merchant Study, July 2011 Retail merchant census: Payment methods accepted, supplier relationships, number customers per day, average transaction size for 4,262 merchants in two areas; Merchant transaction recording: 61 merchants for 4 days each, record every transaction, payment method, size during full trading hours; Customer exit interviews: 466 interviews on payment choices; and Retail merchant interviews: 61 interviews on payment device perceptions and preferences. Kayole 3,489 merchants Kerugoya 773 merchants
Mobile money is the most accepted electronic form of payment, but it is still only accepted by 18% and 20% of merchants in the sample. Many offer financial services on site with credit, layaway, and installment repayments. Payment methods accepted by merchants in census (% merchants) Cash Credit Mobile money Instalments Bank cheque Layaway Uplifting Card Vouchers 20% 18% 20% 15% 48% 53% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Kerugoya Kayole
Neither place is anywhere near cash lite yet. 100% 80% 60% 99% 99% There were 5 card and 4 mobile money transactions out of 6,382 total transactions recorded in Kerugoya. 40% 20% 0% 0% 0% 0% 0% 1% 0% 0% 0% 0% 1% Cash Cards Mobile Money Credit Bank Check Hire purchase or layaway Kayole Kerugoya 39
40 Average transactions are unsurprisingly higher in value when e-payments are used (but remember small number of e-payments and all were recorded at only one merchant, a hypermarket) Mean Transaction Size by Payment Method (Kayole) Ksh 7000 6000 5000 4000 3000 2000 1000 0 359 5,161 6,152 Cash Card Mobile 40 14% of all transactions above Ksh4000 are done using e-payments
And, average transaction sizes are VERY small. But, if you imagine we could convert all transactions >Ksh500 to electronic, that would mean 12% of transactions by number, but 59% (Kayole) and 53% (Kerugoya) of all transaction value. Share of Transactions by Size > Ksh 3000 Ksh 2000-3000 Ksh 1000-2000 Ksh 500-1000 Ksh 200-500 Less than Ksh200 2% 1% 1% 1% 4% 4% 6% 8% 14% 18% Median transaction sizes were slightly higher in Kerugoya at Ksh100 (compared to Ksh 65 in Kayole) (unweighted) 72% 68% 0% 20% 40% 60% 80% Kayole Kerugoya 41
42 Where is the heavy cash? Here, there are fewer clear low hanging fruit, with few businesses experiencing both large transaction sizes and large gross daily revenues. 250000 100% 200000 150000 100000 50000 0 Avg. Tx Val per Day Share of transactions <Ksh200 80% 60% 40% 20% 0%
43 And, Where are the hot zones for transaction volume? These could be big wins for affecting the payments landscape. Most are in the formal economy. Median Tx Size Last mile 600 500 400 300 200 100 0 Avg. Median Tx size Avg. Number of Tx/day >100 TX/ day 500 400 300 200 100 0 Number of Tx per Day
In this study, 95% of customers interviewed had M-PESA accounts and use them fairly often just not for merchant-level purchases. 100% 75% 50% 25% 0% Top reported uses of M-PESA by customers: Money transfer Electricity bill School fees Debt repayment M-PESA Debit/ATM Card Cash
Non-cash payments methods are actually only making major inroads few types of transactions. Payments of salaried workers moving to bank transfers. Wages G2P payments International Remittances M-PESA capturing person to person domestic remittances. P2P Cash merchants School fees (especially among non-poor) moving towards M-PESA or bank transfers Bill payments Retail shops Online merchants Non-local, high value stock purchases are moving towards cheques. But 35% non-local (and 100% local) still in cash. Distributors Wholesalers Manufacturers 45
46 Moving towards cash lite requires prevalent usage of an electronic payment mechanism, but M-PESA is currently ill-suited for this purpose. Consumers resistant to use, because: Bear high fee to send and cover merchant withdrawal cost; Most consumer transactions are small, <KES 100 making M-PESA fees appear a high % of transaction; and Sending an M-PESA payment takes more time than cash. Merchants resistant to accept more widely, because: Inconvenient because they make purchases and payments in cash; Fear of fraud & client reversals ; Usually means paying later ; Network delays & outages; Increased transaction time; and Fear of tracking, tax enforcement.
47 Why hasn t MPESA been adopted, even for free transactions, like Safaricom airtime purchases? Budgeting/Self-control. Most of our respondents purchase one KES 20 scratch card every day. They say that purchasing the scratch card and not developing a habit of buying airtime over M-PESA enforces self-control. Since they would have to go out again to buy a new scratch card, they are less tempted to overspend on airtime, a temptation good for many poor households.
48 So, M-PESA scores high marks, then, for Accessibility Testability Consumer recourse But since it s just a payment mechanism, it comes up short on internal mechanisms to help users exercise spending discipline
49 Consumer drivers for payment device ADOPTION and USAGE are not much different from those identified statistically in the USA. Kenya Assistance as a budgeting/record keeping tool Cost Security Convenience Reliability United States Setup (Adoption) Assistance as record keeping tool (Adoption) Cost (Usage) Security (Usage) Convenience (Usage) Providers agonize over consumer readiness to adopt and use, but there is little evidence that these choices are driven primarily by demographics (education, income, etc.), and much more suggesting that it is a function of perceived product features.
Preventing these payment types from becoming silos means proactively going beyond P2P, focusing first on dimensions with biggest immediate opportunities. Very strong existing case. Strong existing case; unexploited opportunity. G2P B2P P2P P2B Payment Silos B2B G2B Viable case, but very little progress to date Integrated Payments Space P2G B2G Weak immediate case.
Getting to this asks a lot of private providers: Big opportunity, but also big risk especially as they move into new payment domains Integrated Payments Space Providers know they need pervasive high volume payments to make business model work, but uncertainty about consumer response: 1) Speed of adoption and behavior change? 2) Price elasticity of demand? What is happening now that M-PESA tariffs have changed? 3) What are the possible spillovers across payment domains?
There are opportunities for public entities to help accelerate Kenya s path to i-fi by decreasing some of that uncertainty with leading and monitoring research. Building a Cash- Lite Research Agenda for Kenya Obj.1: Reduce risks to private providers by investing in information generation around key design issues. Obj. 2: Highlight opportunities to introduce better e- payment mechanisms in payment dimensions where demand is already ripe. Obj. 3: Encourage e- payments to grow in a socially-inclusive way. Obj. 4: Monitor progress towards an inclusive cash lite economy.
53 Julie Zollmann Bankable Frontier Associates jzollmann@bankablefrontier.com +254 700 581 871