Governments around the world have within their hands a powerful catalyst to promote financial inclusion – their own payment interaction with low-income households. Indeed, government is frequently the largest micro-payer and bill payer in the country. Delivering government payments electronically will not only connect low-income households to an electronic grid, but the government itself will derive sizable benefits from automating these flows.
A recent study by McKinsey, which was commissioned by the Bill & Melinda Gates Foundation, explores the inefficiencies in India’s government payment systems and estimates the monetary and non-monetary benefits of making all of India’s government payment flows electronic. The study finds that automating these payments could save the Indian government US $22.4 billion per year, equal to nearly 8 percent of the total payment flows between the government and its citizens.
To put these benefits into perspective, these savings could alternatively reduce India’s fiscal deficit by 20 percent, boost India’s welfare spending by 25 percent or fund the entire cost of India’s Food Security Act (estimated at $12 billion annually) for nearly two years. Connecting all Indian households to an e-payment system will also generate large but less measurable benefits by increasing uptake of social welfare programs, making it easier to monitor suspicious transaction flows and boosting tax collection.”¨
Sources of Inefficiencies in Government Payments
In FY2008-2009, payments between the Indian government and individual households – in the form of social welfare payments, salaries, tax receipts, subsidies and other transactions – amounted to $296 billion, or $250 per capita (almost one-third of India’s per capita income). A substantial portion of these flows falls into the wrong hands or is spent administering these payments through costly cash- and check-based systems. The inefficiencies in government payments can be classified into three types:
Leakages (75-80 percent of total losses). These occur due to the diversion of benefits to unintended individuals or groups. The government incurs a loss when it makes payments to people who are not the intended recipients, and beneficiaries lose when they fail to receive their full entitlement because some or all of it is extracted by intermediaries, often illegally. A government investigation of India’s National Rural Employee Guarantee Scheme (NREGS)  in a district in Tamil Nadu, for example, found that the beneficiary rolls contained names of deceased individuals and absentee workers.  Another investigation of over-invoicing for works programs found that nearly $900,000 was embezzled in the purchase of materials for 18 local governments. Misrepresentation of man-hours in public works programs is also prevalent. The size and scope of these abuses would be reduced if each of these payments were made electronically using a secure verification system.
Transaction costs (15-20 percent of total losses). These comprise the higher cost of making payments manually (using cash or checks) compared to electronic processing. Time and effort are spent transferring checks between central-, district- and village-level distributors and searching for misplaced checks, often resulting in delayed payments into the necessary accounts. One director of India’s flagship primary and secondary education program, Sarva Shiksha Abhiyan (SSA), pointed out that, prior to the automation of payment flows in his district, he would spend one to two full days each month signing checks for vendors, contractors, teachers and schools. While automating these payments required a fair amount of time upfront gathering bank account information and implementing the process, once the system was fully automated, the director could devote more of his time to overseeing the program.
Administrative and overhead costs (5-10 percent of total losses). Most payment flows are accompanied by an audit or reconciliation process, and the use of manual payment mechanisms substantially increases these burdens. The state of Maharashtra, for example, audits every publically-owned grocery store (called “fair price shops”) in the state twice a year. Each audit involves examining thousands of individual handwritten entries in the shop register and then auditing and reconciling those entries. These functions could be performed faster and at lower cost were all government disbursements to and from these stores made electronic. Similarly, welfare budget planning is especially cumbersome when the data to inform this analysis (such as regional and local disbursements for the previous year) are maintained in manual records, making data aggregation and analysis costly and complex.
Who Bears the Costs of Payment Inefficiency?
McKinsey estimates that automating these payments could reduce inefficiencies and leakages endemic to current payment flows and save the Indian government up to $22.4 billion ($250 per capita) in total annual payments. The bulk of savings ($18.3 billion) would come from welfare schemes, where payment inefficiencies in the form of leakages and administrative costs reach 30 percent or higher (see table below). The potential for savings is as high as 58 percent for India’s food subsidy scheme, but otherwise varies from 10 to 30 percent across other programs. Of the $22.4 billion in savings, roughly $15.8 billion would accrue to the government, $6 billion to individual beneficiaries and $0.6 billion to intermediaries, such as banks, contractors and suppliers.
Just as important as monetary benefits, the government will reap considerable indirect benefits from connecting households to an e-payment system. A transparent payment trail (coupled with a robust identity verification system) will make it considerably easier for the Reserve Bank of India (RBI) and the India’s Home Ministry to monitor suspicious financial transactions, reducing instances of misappropriations or money laundering. Similarly, as cash transactions decline, so will the size of India’s cash-based informal economy, leading to more money in legitimate systems and improved public revenues through increased tax collections for the Ministry of Finance. Even a 10 percent reduction in India’s informal economy could result in an estimated $11 to $13 billion tax windfall for the nation. 
Implementing e-payments will also reduce stress on government administration, as it ultimately leads to reduced corruption and improved law and order. As audit reports uncover multiple transgressions, a significant workload falls on the state policy and judicial machinery to prosecute these cases. E-payments would not only reduce the number of such cases but would also reduce the time required to process each case, as the systems underpinning these transactions will be far more transparent.
Service institutions, like clinics and schools, could also benefit from an e-payment system. Take, for example, the Indian government’s maternity program, Janani Suraksha Yojan (JSY). Despite significant government outreach, institutional child births in India account for only 44 percent of total births. The JSY program provides rural women with financial incentives to have safe institutional childbirths. However, take-up of JSY has been spotty in part due to inefficiencies in benefit payouts. Some reports suggest that it takes anywhere from three weeks to a year to receive a JSY payment.  More timely delivery of these payments will likely induce greater uptake of the JSY program and others like it across India. Government payments could also serve as a vehicle to extend financial inclusion to low-income households. The Government of India has taken a big first step in this regard by encouraging social welfare ministries to deposit payments electronically into entry level “no-frills” bank accounts. This gives low-income households a safe place to store and access their funds. If funds are delivered into the account electronically, outbound domestic remittances, bill payments and other transactions could also be enabled, thus giving welfare recipients both monetary storage and transactional functionality. 
What Will It Take To Build An E-Payment Platform That Reaches Everyone?
It’s a grand vision, but what is needed to make it happen? Connecting all Indian households to an electronic grid will require a concerted effort by government and commercial stakeholders to get three elements right: 1) building the necessary infrastructure to capture government flows electronically; 2) solving the last-mile challenge of getting these payments to the end customer; and 3) finding a business model that incentivizes banks to serve low-income customers. We discuss each of these elements in turn.
Creating the necessary infrastructure. Fully automating these payment flows will require significant government investment in infrastructure. All government departments and agencies will have to be fully networked to ensure that all information transfer is electronic. This includes investment in basic computer and software application infrastructure as well as broadband connectivity to a centralized network. Making e-payments a convenient service for citizens will also require installing transaction points in every village and every urban neighborhood. To ensure payment reliability, a tamper-proof identity authentication infrastructure will have to be installed, especially at payment points that serve rural sections of India with low-income, illiterate households. This will enable the government to direct all payments into an individual bank account that can be accessed only through a secure authentication device.  As shown in the exhibit below, we estimate that this will entail a one-time cost of $13 billion to $15 billion that will pay for itself in a year against potential savings of $22.4 billion annually.
Solving the last-mile problem. Automating government payments and directing these flows into an account is not enough. Low-income individuals will not use the account (other than to withdraw their benefits in full) if they must travel long distances and wait in long lines to access it. Take, for example, a recent study commissioned by the RBI that analyzed the financial access profile of a district that had been declared 100 percent financially included.  The study found that 87 percent of accounts (many of which were opened by customers to receive NREGS payments) were dormant 12 months after account opening, largely because banks had made little effort to establish transaction outlets close to remote households.
Banks are reluctant to build these outlets due to the high fixed costs of deploying branches and ATMs and staffing bank employees in low-income and rural areas. These fixed costs put banks on a slow path to universal financial inclusion and mean that most government payments will “land” in accounts that are inconvenient for low-income individuals to use on a day-to-day basis. Banks should seek to minimize their fixed costs by embracing branchless banking models that take small-value transactions out of banking halls and into local retail shops, where merchants – such as airtime vendors, petrol station attendants and shopkeepers – register new accounts and convert customers’ cash into electronic value and vice versa. Branchless banking turns banks’ fixed costs (branches, ATMs, staff salaries) into variable costs (paying merchant commissions based on transaction volume). This dramatically reduces the revenue threshold needed to establish a viable transactional outlet. In September 2010, the RBI took a major step to enable banks to deploy branchless banking networks by allowing them to engage for-profit companies with extensive distribution networks, such as mobile operators and petrol companies, to set up and manage cash-in/cash-out networks on their behalf.
Finding a business model that incentivizes banks to service low-income customers. While branchless banking may make it easier for banks to expand their physical footprint, banks will not establish and manage transactional outlets in areas with low-income individuals unless they can find a revenue model that makes it profitable to serve low-income customers. To do this, banks must rework their business models away from those things that low-income individuals don’t do much of (leave large balances, take on lots of credit) and embrace what they do want to do abundantly: make frequent transactions to help make ends meet on a daily, weekly, monthly or seasonal basis.  Banks might make it easy for low-income households to park money when they’ve had a good day at the market, save up for school fees that are due next month, save down on the crop proceeds they got three months ago or send money to support relatives in need. This means shifting from a float-based to a transaction-based pricing model, so that banks make money every time customers do something. That is a model that has served mobile operators well: no commitments, no monthly fees, no minimum spend – but you pay each time you call or send a message. The key principle here is to make transactions profitable. If each transaction is profitable, then by definition, every customer is profitable. This financial certainty should push banks to adopt mass marketing approaches, and that is what it will take to expand transactional points to low-income and rural areas. Only then will government payments provide these households a genuine entry point into the formal financial system.