Credit Cards: Optimism and Equal Caution (#74)
A dive into how the RBI's actions and the slowdown in retail credit have influenced the credit card industry
Welcome to the 74th issue of Unit Economics. In this article, I share my thoughts on the current state of credit cards in India. More particularly, I talk about the impact that the retail credit slowdown and the RBI’s actions have had on the economics, partnerships, and processes around credit cards. Dive in!
Credit cards are on track to drive over ₹20 lakh crore worth of spends (>$235 Bn) in 2024. For banks, on traditional accounting of interest, fee, and interchange income, this translates to a ₹1 lakh crore+ annualised revenue pool (>$11 Bn).
Few years ago, banks had suspected that UPI, which - in only 8 years - has grown to power 12X the credit card spends, would chip away at this large pool.
But, even at today’s scale, credit card spends are growing at >15% YoY. And there is much optimism that the pace would continue for a good while.
I share the optimism, but also notice a lot of new ifs and buts today than, say, three years ago. Particularly on the intent and ability for the banks to grow their credit card base at the same pace.
But before I get there, the larger context of retail credit needs to be well understood.
The retail credit slowdown
Retail credit in India has gone through an unusual transition over the last couple of years. Take, for instance, home and personal loans, which account for ~55% of all retail outstandings. In FY24, the growth rate of home loans halved year-on-year. Personal loans, on a path for 25-30% year-on-year growth, fell off even worse.
This change in momentum was sudden. But not completely unexpected. Over the last three years, the RBI has continuously called out two large systematic risks in retail credit:
an unusually high credit–deposit (CD) ratio,
increasing delinquencies in specific segments, particularly in microfinance portfolios
The high CD ratio, data suggests, is due to a structural shift in household preferences post-COVID. Where households began spending more on asset purchases and discretionary consumption, and to finance them, (a) took more loans and (b) diverted a higher proportion of their savings away from bank deposits. Putting pressure from both sides.
The phase was also marked by collective exuberance from the lenders, many of whom - especially those non-bank - adopted a growth-at-all-costs approach. Delinquencies followed across portfolios. From small-ticket unsecured loans to credit cards, education loans, and vehicle loans.
The RBI knew that it had to intervene, and it foreshadowed the intent in early-2022 through continuous (a) increases in repo rate [i.e. the rate of borrowing from the central bank] and (b) guidance to regulated entities against any exuberance. In fact, between May-22 and Feb-23, it increased the repo rate by 250 bps. This was the highest absolute change in the rate within a 12-month period since the global financial crisis (2008).
However, despite the banks passing the rate increases to the customers, the consumption patterns and the demand for retail credit sustained. The soft guidances were also turning out ineffective in changing lending growth practices of the regulated entities.
The RBI grew impatient. Very soon, it resorted to direct force to bring down the retail credit supply. And it did that through two methods.
It formalised and stopped the many large-scale, murky retail lending practices, introducing (a) digital lending guidelines, (b) FLDG caps, and (c) higher risk weights for consumer credit, among other smaller supply-limiting regulations.
It strictly enforced existing guidelines around retail credit through (a) regular audits and (b) harsh penalties on the slightest of negligences.
The impact of the crackdown (eg: 1, 2, 3, etc.) was immediate. The growth in loan originations subsided across all categories of retail credit. And the spike in defaults lately suggests that the interventions may have been just in time too (though we may question the manner).
Amidst all the noise, how did credit cards fare?
The credit card spends have held up well, showing a downward impact of not more than 4-5 pct points through the year.
The lower impact on credit card spends versus other forms of retail credit is a result of three factors:
Limited influence of new, external actors
Unlike unsecured loans, where newer and less trusted NBFCs / MFIs drove as much as ~60% of new volumes in FY24, credit card issuance and management remain bound to scheduled commercial banks, as dictated by the RBI.
While that does not take away the sting from the central bank's lens, as HDFC, Kotak, and Bank of Baroda experienced, it does suggest the existence of (a) more sustainable offline and online channels for distribution and engagement built over decades and (b) better compliance practices. Both of which help mitigate the industry-wide impact of any regulatory action.
Larger portfolio dependence on existing, carded customers
In the last two years, the average number of credit cards per carded customer has gone up ~30% from 1.7 to 2.2 (est.), while the spends per card have increased from ~12.5K to 15K (~20%) in the same period. That is, there has been sufficient demand from existing carded customers for (a) acquiring more credit cards and (b) increasing usage of the cards they hold.
This is again unlike unsecured loans, where there is a large portfolio dependence on new originations to increase the total outstandings.
Growing, positive perception of credit cards as a transaction tool
The demand for credit cards remains strong, and faces little risk due to
shifting preferences in tier 1 and 2 cities towards more discretionary expenses and household credit,
continued marketing for credit cards by NPCI and the top banks, increasing the awareness and helping fight the taboo in tier 2-3 cities, and
differentiated positioning of the credit cards, as a payment method that can be both (a) rewarding and (b) helpful in managing cash-flows
Unfortunately, the mitigators sustaining the growth in credit card spends do not translate as well elsewhere. The absolute new originations of credit cards, for instance, fell more than 50% YoY in 2024.
Banks have, no doubt, grown more risk-averse with (a) the increase in risk-weights on credit cards, (b) the urban spending slowdown, (c) the rise in delinquencies (though some estimates seem quite off) and (d) the fear of regulatory action. And it would be easy to attribute the drop to all things external.
Yet, parallelly, I notice five growing stressors internal to the banks that are adding to the change in momentum. I detail each below.
KYCs have hit a U-turn
No other master directions from the RBI have likely been revised as often as the directions for Know Your Customer (KYC), 2016. If you go through the revisions, you could divide them into three neat phases.
The early 2016-19 phase was focused on setting the foundations for digital KYC. It oversaw the introduction of OTP-based Aadhaar verification and the Central-KYC registry.
Then, with the baseline defined, the 2019-2022 phase was marked by an exponential growth in the adoption of all things digital. This phase was driven primarily by tech improvements in and regulatory acceptance of (a) electronic Aadhaar-PAN verification, (b) Video identification (V-CIP), (c) liveness checks, and (d) transfers of CKYCR records.
By the end of it, the regulators and TSPs had managed to make the full-KYC process completely digital, which made the average KYC significantly cheaper and faster.
The tech-enablement also allowed fintech companies more control of the customer experience. End-to-end digital KYC became the norm. Even for the issuance of credit cards, and especially for the co-brand partnerships.
It was assumed that the process would only get more tech-dependent over time, with AI and ML the common buzzwords.
However, over the last couple of years, during 2022-24, the improvements in and the adoption of KYC-related tech appears to have stalled.
Why? As is often common in India, the loopholes in the Digital KYC process were exploited much faster by fraudsters than the banks or TSPs could fix them. Cyber frauds became common, customer complaints rose, and the regulatory mandates followed.
Over time, the regulator and the commercial banks started taking a more conservative view of Digital KYC. And lately, the steps appear to be all backwards.
For instance, KYCs of credit cards today see a higher proportion of customers being categorised as risky, unless some prior form of employment or personal verification is performed. To those that are categorised as such, a field investigation or contact point verification becomes mandatory for KYC completion. Making the entire process offline, cumbersome, and costly.
Further, the RBI has taken a very aggressive approach towards all identity and address verifications, putting in highly restrictive AML practices. Take, for instance, the new mandate around necessarily establishing that the applicant’s Aadhaar and PAN are linked, to limit the identity fraud which was happening on account of new, fake PANs. This, RBI suggests, could be stopped if each PAN was linked to its owner’s Aadhaar.
However, while it may save against the small section of fraudulent customers, in doing so - a higher number of genuine customers will likely get de-banked due to their lack of awareness and fulfilment of the mandate.
The frequency and ferocity of such changes to the KYC has increased over the last year, and while the regulator’s intent is well-appreciated, the banks have increasingly erred towards guarding against regulatory risk more than the customer risk on their KYC processes today, impacting their approval rates.
Underwriting policies have become more exclusionary
With the RBI’s interventions and higher defaults on retail credit, the banks’ underwriters have tightened their policies on unsecured credit, publicly claiming to have put drastically higher caps on credit scores, cuts on credit limits, and limits to other rules too.
This has translated directly to a lower credit card underwriting approval rate and an even lower share of New-to-Credit customers (only 12%) in the approved-mix, increasing the costs for new credit card issuances.
An increase in such costs also has an exclusionary second-order effect. It shifts the underwriter’s focus towards safe existing, carded customers. And at a portfolio level, it takes away the priorities away from any form of financial inclusion.
Credit card economics are in distress
Taking SBI Cards’ financials as suggestive of the broader market economics, the banks have a lot to be concerned about.
For instance, the Return on Average Assets (ROAA), traditionally between 5-6%, fell to 4.7% in FY24 and has since come down 220 bps to 2.5% in the Q2 of FY25.
The ~40 to 50% fall in ROAA is largely driven by the increase in credit costs, which as a proportion of average assets, have increased from 5.4% in FY23 to ~8% in H1 FY25. Roughly 85% of this increase can be attributed to the increase in gross write-offs. Although the provisions are not translating to higher Expected Credit Losses % (ECL) yet, as per banks’ measures.
The increase in write-offs is despite the proportion of revolvers holding at 24-25% and EMIs at 37-38% across the portfolio. Implying that more than the intent, it is likely a change in conditions related to the customer’s ability to repay, which is making the revolver-to-defaulter funnel broader.
Bankers signal that it might be due to the poorer macro conditions for the middle class, which is struggling to accumulate savings or to access personal loans for paying off credit card dues. The latest GDP numbers support the thesis.
Similarly, there are signs of stress in revenue components as well. Interchange income as a portion of average assets has gradually come down from ~18% to 15%. This is likely due to the increasing proportion of RuPay / UPI in the credit card portfolio, lowering the blended interchange fee.
On the other hand, interest income has continued to be at similar levels for the last three years, but with (a) the RBI’s stick on high APRs and (b) the rising cost of funds (CoF), there are pressures against its sustainability as well.
Credit card reward programs have weakened
Carded customers have become accustomed to devaluations in reward programs today (look at their frequency on Reddit). This was expected, with the hands of most banks forced by the distress in the card P&Ls as earlier highlighted. Further, the product costs for lounge access has also increased, making it tough to retain another key value proposition for cards.
The nature of stress on the credit card economics suggests that the situation on reward programs will directionally remain as such for another couple of years. However, due to this trend, I would bet on the following second-order effects to grow common:
Banks would increasingly seek large-scale co-brand-partners to provide brand-led value propositions. Cutting deals similar to the Swiggy, Flipkart, or Amazon co-brand cards. This would help them (a) leverage the merchant’s distribution to lower their acquisition costs, and (b) offer meaningful rewards, without high product costs, both of which are big pains for them.
Banks will increase the incidence of milestone-based rewards, and limit waivers and rewards towards only engaged, monetizable customers. For instance, banks are already increasingly introducing the annual fee and fee for the plastic, both of which were commonly absent on a lot of mass-market cards over the last few years. And then allowing waivers of the fee only on reaching certain spend milestones.
Average number of new cards per existing, carded customer, will increase much slowly. Such customers are likely be less enticed with newer cards due to the weaker propositions. This will also push banks to manage their NTCC funnels better.
Power dynamics in co-branding partnerships are skewed towards banks
Lastly, unless you are a merchant with tens of millions of customers, the power dynamics of co-branding arrangements have shifted against you - with the small-scale fintech companies most adversely impacted.
The RBI has intentionally and aggressively taken away the control from the co-brand partners, limiting their ability (a) to partake in risk-sharing via FLDG caps and (b) to cross-sell or provide superior experience to the acquired customers using credit card data.
This has limited the value of a co-brand partner largely to a marketer / distributor of the cards, as intended. Consequently, the commercials of such arrangements have become restrictive to the partners, until they can leverage some existing, large distribution in their negotiations.
Further, the ability of co-brand partners to handle card onboardings has been called into question multiple times in the last couple of years, with frauds registered on fairly renowned applications. The corresponding penalties from the RBI on the regulated bank issuers have deteriorated the trust on co-brand partners. This has forced banks to take drastic actions on the customers acquired via these partnerships, commonly cutting down credit limits and underwriting approval rates. It has also lowered the leverage with co-brand partners to negotiate a better customer experience against the banks’ risk teams, which used to be a non-negotiable a few years ago.
A direct consequence of the changing co-branding dynamics is that there appears to be little merit in cornering the business model of a fintech around credit cards today, unless large distribution already exists. More directly, I would suggest that it would be unwise to start a Kiwi or OneCard in today’s environment.
P.S. Expect, however, the rollouts of co-brand cards to continue across the ecosystem, with the partnership push from the NPCI on RuPay. Although, the proof of success will rest in the scale of such programs.
Final few thoughts
Banks have previously navigated through similar credit card portfolio distress and should be able to come out better in another year or two. The many shifts, however, in the KYC process, underwriting rules, rewards propositions, and the nature of co-branding partnerships are likely to stay much longer.
With that, we should expect the growth in new originations to remain a bit mellow as well. The focus for business teams will then continue to remain on (1) maximising engagement of existing customers, and (2) figuring out new ways to cater to new, low-touch segments.
On (1), I expect three, already ongoing developments to continue pace: (a) banks will push EMIs on credit card transactions further for transactors, to increase interest income, (b) NPCI will continue to push banks and fintech companies credit cards on UPI, which should migrate small-ticket expenses towards credit cards for at least 20-30% of the carded base, and (c) banks will strengthen their reward value propositions by building merchant marketplaces, using subvention to make the card propositions more enticing.
On (2), the ecosystem will hope that (a) the bet on Secured Cards plays out well over the next decade, especially with the improvements in tech for digitally opening and linking FDs to credit cards today, and (b) some of the vertical niches grow large enough - say, around foreign travel or e-commerce - to allow standalone card P&Ls to be built on them.
P.S. I have recently joined a very talented team at super.money to build credit cards, and will be closely observing the market for the next few years.
Meanwhile, 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!