How to kill cash

There is a general sense of reluctance when it comes to electronic payments in the plains of offline retail in India.

Merchants aren’t interested in accepting them, banks aren’t interested in convincing any other than the biggest merchants to accept card payments and most users are not interested in using digital payments for most things.

The government, eager to drive cashless payments, has tried everything. Slashed the fee merchants had to pay banks for accepting payments, gave banks targets to hook more merchants with point of sale terminals to accept digital payments and is looking at pushing low-cost solution like QR codes.

But all of this becomes a Sisyphean task, because of fundamentally misaligned incentives for all stakeholders in the digital payments system. But what if merchants who today are pushed and prodded to accept digital payments actually want to accept digital payments?

And that starts with fixing the card payments before we get to pushing something like the payments poster child UPI (short for Unified Payments Interface, a real-time mobile payments system).


In Jan 2019, UPI hit a record 672.75 million transactions. The National Payments Corporation of India, which manages UPI, doesn’t split the data, but let’s assume that half of these volumes are made from individuals to merchants. And more than 90% of these merchant payments are online—for recharges, bill payments, e-commerce, etc.

The vast majority of commerce is still offline—and cash-heavy

Zoom out a bit here. More than 95% of commerce in India is still offline, and the picture is quite different there. It is cash-heavy and limited by point of sale (PoS) devices present only in the large cities.

You may ask why bother about card-based PoS payments instead of faster, better, and cheaper (UPI-based) QR-code solutions that consumers scan with a smartphone. Aren’t they an obvious solution to reaching consumers and merchants at scale? Not so fast.

The real picture is sobering. Of India’s 390 million Internet users, about 160 million transact online, said a recent report by Google, Omidyar Network and Bain & Company. More importantly, 54 million of the 160 million stopped transacting online after their first purchase—presumably because of a poor interface and other challenges.

Add to this the fact that the average user already has 51 apps on their phone. For QR code solutions to work at scale, consumers will need a compelling enough reason and relevance to download that 52nd app. Not to mention having a stable enough connection to complete the payment instantly at the point of sale. Both of which are optimistic, at best. It’s perhaps more realistic to persuade a consumer to pull out a piece of plastic that they already use at an ATM and present it to the merchant.

The Reserve Bank of India (RBI)’s credit- and debit-card usage data for November 2018 shows that credit cards are a sixth of the market by instruments. And they represent roughly half the card spending at PoS devices.

The Payments Council of India, an industry body, wants more credit cards in the hands of Indians. Their suggestion to the newly formed, Nandan Nilekani-headed digital payments committee was to now allow non-banking financial companies to issue credit cards. Though well-intentioned, it will simply add another payment instrument in the hands of people who already have a debit card (Indian banks have issued close to a billion debit cards).

Instead of adding new payment instruments in the hands of the same consumers, what can the financial industry do to get them to use what they already have—debit cards—for retail payments?

People know how to use debit cards, but don’t find enough PoS terminals—so they end up using cash. The argument then is that we need to incentivise the merchant and the acquirer, and can do so by smartly tweaking the economics of a payment transaction—rather than subsidies that end up distorting the market. But how did we get here?

A long and winding road

In an effort to make debit card payments affordable to small merchants, the RBI has capped the merchant discount rate—the fee that merchants pay banks for providing the service—at 0.4-0.9% of the transaction value.

The input costs for the acquiring bank is the interchange fee that the card networks like Visa and Mastercard charge. Rs 0.15 (0.2 cents) for transactions under Rs 2,000 and Rs 1.5 (2 cents) for anything higher. So an average debit card PoS transaction of Rs 1,435 will see an average MDR of 0.5% being charged by the acquiring bank. (We need to take an average, since the MDR caps depend on the merchant’s annual turnover, not individual transaction size.) So an acquirer’s net income is about Rs 7 per transaction after netting out Visa/Mastercard’s interchange.

And across an average of 107 debit card transactions per PoS device per month, that’s paltry average revenue per user of Rs 749 per device per month. That for a business which has high operational costs—feet on streets to deploy and service PoS terminals, customer service operations, etc. 

Given these economics, banks limit themselves to trying to acquire merchants who can drive a minimum threshold of transactions. The vast majority of smaller merchants are not viable. No wonder there are still just 3.5 million PoS terminals in a market of 14 million retail outlets.

Even within the currently “terminalised” merchants, competition in the PoS (i.e. merchant-acquiring) business is intense, margins wafer thin, and differentiation hard to achieve.

Most acquiring is still done by banks; independent firms such as Mswipe, Ezetap and MobiSwipe, who seek to expand the market, have not been attractive for investors, raising a combined $273.1 million over the two years since demonetisation, according to startup database Tracxn. Even large acquirers such as Pine Labs, which claims to operate 300,000 PoS terminals, seem to be de-risking their business by buying issuing side businesses which are arguably more profitable.

Together they account for about 500,000 out of an installed base of 3.5 million PoS devices. And Reliance’s pending entry into the PoS business is going to make the investment case in standalone PoS vendors even more difficult going forward.

Usage trends

Every month, each of India’s more than 200,000 ATMs see over 4,000 transactions from debit cards, while each of the 3.5 million PoS devices see 150 transactions (across debit and credit cards)

For context, debit cards were used at ATMs 0.85 times a month, and at PoS terminals only about 0.38 times a month. 

What goes on when you consider the economics of a payment transaction is this: When you use your credit card, the issuing bank (the one that issued the card) pays the acquiring bank (which got the merchant to sign up for a PoS terminal) within three days. But it gives users between 21 and 51 days to pay that amount back, depending on your billing cycle. So the bank incurs a funding cost of about 35 days.

If you choose to roll over the credit, there’s a risk to recoveries and provisioning costs. So in all, the costs sort of justify the issuing bank charging the acquiring bank an interchange fee—usually 1.8%.

But what happens when you use your debit card?

The issuer is sitting on good funds—yours—which they simply move to another bank’s account with zero funding risk. Consider two scenarios.

One, you use Bank A’s debit card at Bank B’s ATM. Because Bank B gives you the cash immediately, Bank A pays Bank B a fee of Rs 15 for compensating Bank B’s costs of float and running the ATM. Makes sense, right?

But with PoS transactions, it is the opposite. If you use your Bank A debit card at a merchant acquired by Bank B, Bank B is not only settling with the merchant “T + 1” in industry speak (meaning Bank B pays the merchant after a day), Bank B also pays Bank A an interchange fee!

To cover its costs, Bank B charges the merchant an MDR fee. The issuing bank (Bank A) has zero risk and zero funding costs. Yet they’re earning fee income from a merchant they don’t serve for doing nothing except move money to Bank B over RTGS at a next-to-nothing fee.

Why? An unwillingess to challenge the status quo.

Card networks have rewarded the issuer disproportionately over the acquirer. The split of MDR is about 70% in favour of the issuer bank—a legacy of developed markets that are overwhelmingly credit-driven. In India, where debit cards rule, we simply replicated the model with lower fee levels, hoping that increased volumes would offset lower per-transaction fees. This hasn’t worked. 

Fortunately, a possible solution is logical, practical, and scalable—but needs a will to challenge conventional wisdom.

Remaking the payments house

“Learn the rules like a pro, so you can break them like an artist’”
Pablo Picasso

Good intentions—reduced fees, MDR caps, mass-issuing debit cards—are necessary but insufficient. The payments industry need to go beyond good intentions to developing strong mechanisms to drive the outcomes it needs. That means reinventing, and not tweaking, the existing model. So here are a few ideas:

  1. Reverse the fee structure, make debit card issuers (Bank A) pay acquirers (Bank B)—similar to the ATM story. There’s really no difference in the business model. At, says, Rs 15 per transaction that the acquirer earns, they can sustain merchants at even 50 transactions per month. Suddenly, the long tail of small merchants becomes viable. As for the issuer, though this structure results in higher costs, incentivising the PoS transactions means more data. And that will help better cross-sell to users and merchants alike, leading to more profitable lending revenues. Plus, most banks, especially the large ones, are both issuers and acquirers themselves.
  1. Let a viable business model attract entrepreneurship and risk capital to drive product innovation and scale. To reach even a conservative short-term goal of getting half of India’s retail outlets PoS-enabled needs an additional 3.5 million terminals—doubling the current market size.
  2. Instead of the government subsidising MDR—essentially borrowing from you and me to fund fee income to issuing banks—it should subsidise domestic PoS manufacturing, bringing down hardware costs, developing intellectual property and creating jobs.
  3. No MDR for merchants—they get paid Rs 100 for every Rs 100 worth of goods sold, and receive good funds in their bank accounts the same day. So merchants end up actually asking for—instead of simply accepting—card payments.
  4. Banks use this wealth of transaction and settlement data to offer merchants working capital credit and other stickier or more profitable products. Imagine the network effects of doubling the base of small retailers whose transaction data can drive data-led lending at huge scale.
  5. And then UPI 2.0 comes along and further lowers costs by leveraging its enormous scale.

It sounds simplistic. And maybe wiser people have perhaps thought of this before, but gave up in the face of incumbent banks blocking change. Years of accumulated systems and processes are behind this. Retail banks’ internal metrics treat issuing and acquiring as separate P&L (profit and loss) divisions, and the two rarely talk to each other. Large banks have traditionally focused on growing cheap CASA (current account and savings account) balances—change how debit card payments work, and suddenly you’re changing the economics of the entire business. And finally, you have legacy IT systems that are hardwired to charge transaction fees in particular ways.

But today we have an excellent set of overlapping factors—growing consumption demand, better payments infrastructure, multiple players capable of sustained investments and a regulator open to new ideas.

So if not now, then when?

Leave a Comment