Credit on UPI: Early signals and the next phase (#73)

Expect more patience, competition, and traditional forms of experimentation, amid slowdown and industry-wide challenges

Aug 26, 2024

Welcome to the 73rd issue of Unit Economics.

I have often written on Credit on UPI (May-22, Apr-23, and Dec-23), and with optimism. Much of the ecosystem felt similarly initially. However, over the last two years, many early assumptions have been tested and answered. And we today realise some general truths about where Credit on UPI might be heading.

In this article, I pause and take stock of how it has progressed on its promise, and share thoughts on why it is entering a new phase of more calm and grounded experimentation for the next few years. Dive in!

The RBI first proposed linking Credit Cards on UPI in Jun-22. The NPCI, which manages both the RuPay card network and UPI, then issued the operating circular for the linking of RuPay Credit Cards in Oct-22. Within four months, in Feb-23, the HDFC bank became the first bank to go live on UPI with its RuPay Credit Cards.

These were significant, positive developments in a period filled with scepticism (due to regulations around credit lines on wallets and digital lending).

The early reception around Credit Cards on UPI was a little mixed in the industry. There was a lot of optimism, mostly from fintech companies and VCs, and some fear, primarily from banks. Let me explain each of the two below.

The opportunity for Credit on UPI

There was optimism because the market opportunity was huge. In credit cards, you already had a product with

a large market gap (<5% penetration),
a fast-growing market (>15% CAGR), and
a proven business model that could make money at scale (~5% RoA)

But add to it the scale of UPI and you now potentially had,

38 times more offline merchant acceptance points (340Mn+ QRs vs 9Mn POS devices),
6-7 times the transacting, habitual user base (350Mn on UPI vs 50Mn on CCs), and
access to alternative data for underwriting through UPI transactions

UPI’s scale and ingenuity suggested many positive second-order effects for credit cards: (a) higher average monthly spends and transaction volumes per card by migrating the smaller-ticket UPI P2M transactions to credit cards, (b) lower costs of card distribution by accessing the customers of existing UPI applications, and (c) higher penetration of credit cards to the unserved but credit-eligible customers (160Mn+), on the assumption of a better-quality of alternate data.

Now, if the prophecy were to come true, then the ecosystem expected two large-scale shifts: (a) higher % of credit card origination through fintech platforms instead of the costly banking channels, and (b) higher % of credit card transactions using cards linked to UPI IDs rather than the direct POS swipe, tap, or bank OTPs.

For the platforms gaining or enabling these shifts, these could create new, large revenue pools. For instance, imagine if - at today’s scale - even 5% of all UPI spends were to migrate to credit cards or other forms of credit on UPI. These 5% spends would be worth more than ₹1 lakh crore (~$12Bn) of monthly spends. For the issuers (banks and fintech companies), assume a reasonable blended take rate of ~1% on these transactions. This would amount to ₹1000 crore (~$120Mn) worth of monthly interchange revenue for the industry, which would be quite significant. And this is despite the assumptions above being quite conservative.

Naturally, the fintech companies and VCs, realising the novelty of the tailwind and the opportunity size, were excited by the promise.

Banks feared lower interchange on credit cards and losing customer ownership

On the other hand, while the opportunity created a lot of hope, it equally caused fear. Fear, because UPI had been a painful experience for many industry participants - particularly the banks and card networks.

UPI had already eroded large revenue pools from consumer payments, from wallets, debit cards, ATMs, etc. by following a very utilitarian approach towards the MDR. They feared that if the same spirit were to be exercised by the RBI and NPCI for Credit on UPI, it could damage even larger revenue pools that existed in consumer credit.

To make matters worse for the banks, UPI is one among the few battles that they have conceded to fintech companies. And the loss of customer ownership has cost them a lot of up-sell opportunities.

They feared, quite rationally, that if the issuance and utilisation of credit cards were to shift towards UPI, the third-party UPI apps would likely continue to have the advantage and the banks would struggle further to maintain customer ownership.

But they knew that, like with UPI, if they could not do anything to stop this change, they would rather be in on the action than wait on the sidelines.

Now, in only two years since the introduction of Credit on UPI, many of its promises and fears have been tested. The early signals point to noticeable fundamental shifts in the card network preferences, business approach towards new UPI apps, credit card origination channels, and credit card usage behaviour.

This is significant since each such signal is likely to dictate the design discussions for Credit Cards on UPI for the next 100Mn consumers. I detail five such noticeable shifts below.

Heavy rush towards UPI TPAPs; More use-case-specific approvals

The market pull for Credit on UPI created a lot of demand for (a) Third Party Application Provider (TPAP) certifications, which enables an application to issue UPI handles and initiate UPI transactions on their app, and (b) for RuPay partnerships, to issue co-branded cards with banks.

The list of fintech companies that have done so is fast-growing and includes:

Fintech companies with large, existing consumer distribution: Navi, Groww, CRED, Tata Neu, and Aditya Birla Capital, among others
New companies focusing on Credit on UPI as its core business: Kiwi, POP, super.money, SalarySe, and Rio
Previously certified UPI TPAPs: Jupiter, Paytm, GPay, and PhonePe

There were rumours that NPCI had more than 50 pending applications for TPAP at one point last year, and has since been very particular about who it gives the business approval to. For new applications, there is a higher prominence of use-case-specific approvals, rather than generic certifications for TPAPs issued earlier. This approach from NPCI also aligns well with the business strategy of many fintech companies today, as highlighted next.

UPI apps focus on viable customer segments, not market capture

While the competition amongst UPI applications earlier was all about the mass-market appeal and market share, the new-age TPAPs are taking a more refreshing approach.

They appear to focus more on driving top-of-wallet behaviour for a focused and defined customer segment of not more than 50Mn customers, allowing more controlled unit economics from Day-1. CRED and SalarySe are good examples of this strategy.

Saying that, driving acquisition and engagement on a new UPI app, along with the management of a TPAP licence, are both costly affairs still. However, the bet is that the approach would be increasingly more viable due to (a) the potential for Credit on UPI to eventually drive revenues through interchange (above certain transaction value) and interest income, and (b) lower CACs due to a more focused approach to distribution.

Large gains for the RuPay network on Credit Cards

With RuPay’s monopoly on Credit Cards on UPI, almost every large private and public bank appears to have introduced RuPay variants of their existing flagship cards or introduced new cards exclusively on RuPay.

This has meant that while RuPay made up ~5% of all new credit cards issued in FY 22-23, the figure has gone up to 30% today. The share of RuPay credit card spends in overall credit cards has also surpassed ~10% as a result. Expect both of these figures to continue rising with the continued preference for RuPay, the introduction of card interoperability, and limited favours for Mastercard and VISA.

Another second-order effect of this is a bottoms-up preference from fintech companies to choose RuPay against other networks, to safeguard their regulatory risks.

Card plastic is losing its value

With a 38X larger merchant acceptance base for UPI QRs than for POS terminals and little-to-no exclusivity on merchants, the physical form factor of a credit card, i.e. the PVC plastic or metal card, is losing its value in both convenience and prestige.

Some estimates suggest that ~20% of all new credit cards issued today are only virtual in form. Expect this to only grow, with the lack of utility and the high costs of issuing a physical card. Alternatively, many issuers will likely start charging an additional fee for the plastic to cover its costs.

Credit cards show more top-of-wallet usage when linked to UPI

Customers using credit cards on UPI seem to be very strongly migrating their low-value UPI merchant transactions from savings accounts to linked credit cards. They also continue to be favouring credit cards for higher-value discretionary spends. As a result, this is leading to the following changes in per-card averages:

Increase in monthly spends by ₹3-4K
Double the volume of transactions, from 6-8 to 15-18
Decrease in average ticket size per transaction by more than 50%
Increase in card usage for daily-use categories, including groceries, department stores, etc.

This is particularly positive for UPI TPAPs, since with higher transaction volumes, there is a greater chance for your co-branded credit card and UPI app to usurp all other payment apps in becoming top-of-wallet for the consumer. For banks, on the other hand, this only heightens their risk of losing revenue opportunities with the customer despite their card issuance.

As a result of the push towards Credit on UPI from the NPCI, issuer banks, and UPI apps, RuPay Credit Cards on UPI have surpassed ₹10,000 crore in monthly value ($1Bn+) today, accounting for ~6% of all credit card spends in the country.

However, despite the seemingly large scale, Credit on UPI has struggled with quite a few challenges in scaling up and, surprisingly, fallen short of where it was expected to be by this time. Three such challenges are most prominent.

High underwriting declines; Underlined by Credit Card slowdown and lack of strong alternative underwriting

Traditional banks have long kept a high-bar for credit card approvals, limiting their access to only the top 50-60 million consumers in India. For large scheduled commercial banks, the bar has been raised further lately with the scrutiny on the pace of consumer credit growth from the RBI. This is reflected in an industry-wide slowdown in credit card issuance, despite the rising customer demand.

Credit on UPI applications have borne brunt of this, with the onboarding conversion rate in single digits for most. To make matters more challenging, banks are yet to find success in the quality of alternative data with the UPI service providers. The NPCI did float a credit-scoring platform that would use UPI transaction data, but little is known of its effectiveness.

For the co-branding partners, the lack of control over underwriting (there are some exceptions to this rule) and the struggles of co-branding regulations have made it a more challenging relationship to manage and dictate. The banks are clearly demanding more patience from the fintech companies and the VCs today than before.

MDR remains unsolved; impacting customer rewards and experience

The RBI and NPCI introduced a tiered MDR structure on RuPay Credit Card transactions over UPI, making transactions valued less than ₹2,000 free for merchants and charging 2% for those above.

This has led to two, unfortunate second-order effects:

Due to zero MDR on low-value transactions, the issuer banks earn no interchange on >20% of their Credit Card spends on UPI. This has made their average earnings on card transactions over UPI lower than those over POS terminals or a payment gateway. As a result, they have had to moderate their rewards on UPI transactions to be lower than on a direct swipe. The NPCI does not seem too happy about this.
The majority of the UPI QR merchants, on the other hand, are taken aback by the standard 2% MDR on high-value transactions and are disabling the acceptance of Credit Cards on UPI. Some estimates put the % of such merchants to be >20% of all merchants on UPI. This includes large merchants, who would otherwise want to negotiate better MDR rates, and small merchants, who do not realise any incremental value of accepting credit cards versus the usual UPI payments. Regardless, this has a direct impact on the customer experience and the issuer’s economics.

This uncertainty around the economics, apart from the risk of losing up-sell opportunities to the customer, is another reason why banks are wary of migrating their existing carded customers or allowing higher approvals through the UPI channel.

Credit line on UPI continues to be restrictive; Banks fear credit card cannibalisation

Credit line on UPI (CLOU) appears to have made little progress since the last time I wrote on the subject. It continues to (a) be limited to pre-approved existing-to-bank (ETB) customers, (b) follow credit terms more commonly adopted by overdrafts, and (c) have few takers amongst scheduled commercial banks (while the SFBs and NBFCs remain outlawed).

The position of scheduled commercial banks on Credit Lines is not without good reason. They (a) fear that credit lines, if interest-free and issued to new-to-bank (NTB) customers, could potentially cannibalise their credit card portfolio, and (b) lack confidence in the ability of Credit Lines to deliver strong, profitable financials - like the credit cards or personal loans do; they will want more confidence on the same over time and find it outside the spirit of the bank to rush without so.

There do seem two quite noticeable developments on the subject, however:

Zeta has developed a complete lifecycle management solution for CLOU and is betting big on a potential hockey-stick growth on them. They seem to have bridged the technical gaps that were apparent.
Navi has, after a long search, partnered with Karnataka Bank to offer their version of Credit Line on UPI. This would likely be the first show of the promise that CLOU holds, but may also come under some regulatory stick - on the assumption of them provisioning a high-FLDG share. Regardless, this would be one of the more exciting launches to watch.

While the long-term optimism around Credit on UPI’s original promise continues to hold, the three challenges above have shaped varied emotions for banks, fintech companies, and investors.

Two years in, as some questions continue to linger, we can also be more confident of certain directions that the industry seems to be heading into. I try to capture some such accepted truths below:

On underwriting, any significant unlock on account of UPI transaction data seems unlikely in the short-to-medium term, especially with the underwriting control with the more conservative scheduled commercial banks. Account aggregator appears to be the best option for alternative data presently, but equally continues to carry low evidence of standalone effectiveness.
The industry seems to be taking a large bet on secured credit cards to tackle the credit card slowdown problem. There is bullishness largely due to (a) the presence of lower-cost infrastructure today for enabling FD-backing of credit cards, and (b) the success of Stable Money in proving that there is consumer appetite for FDs. Many are in line to issue FD-backed-credit cards on UPI soon, and if successful, this could create incredible delta in the Credit on UPI’s success.
Credit Cards on UPI continue to remain an expensive affair, and it will be a long and hard journey for the likes of Kiwi, Rio, and Pop - who are venturing into it (and I might be assuming incorrectly here) without (a) a differentiated customer segment and (b) any underwriting, distribution, or collections advantage.
I believe that the likely winners on Credit on UPI will be those who (a) compartmentalise their app away from any existing brand they might be operating, unless it is a UPI app, (b) have a strong underwriting or distribution advantage, and (c) carefully target a sub- 50Mn customer segment, rather than going mass-market. More in line with the strategies of super.money, CRED, or SalarySe. Amazon’s bet on a separate app and Axio are likely steps in the same direction.
On commercials, Credit Cards on UPI will likely continue with a tiered approach and suggested rates. The ecosystem will also have to live with <90% merchant acceptance for a long time, particularly for small merchants, who will not be comfortable with the 2% MDR. For large merchants, negotiating lower merchant-specific MDRs will likely be an easier problem to solve.

Overall, Credit on UPI appears to be transitioning from an early phase of uncertainty and unfounded optimism to one of more calm experimentation and bank-led growth, having slowly accepted the industry’s limitations. Over the next few years, I intend to be close to the action on Credit on UPI, and will continue sharing my thoughts on the developments!

If you want to share any views or talk casually on the topic, please reach out to me by email or LinkedIn. Until next time!

Unit Economics