Ant: Share of profit != share of value
As in fintech as a whole, for Ant payments and credit are the easy, but limited, wins — it’s whole-balance-sheet engagement that makes the difference
One of the seeds for Jupiter was planted a few years ago when I saw the way friends in China reacted to their Yu’e Bao account in AliPay. They would open it and just grin. “I know it sounds silly,” a friend said, “but I just feel an emotional connection to the interest growing”.
I have no view on Ant Financial’s valuation. But Ant encompasses almost all of fintech, at unrivalled scale in each vertical. So understanding how it generates value may say quite a lot about fintech as a whole. There, I believe analyses that begin from the dominant profit shares of payments and credit miss a crucial point.
Payments and credit print cash for now, but they are limited on their own. It is the savings, investment and insurance segments that create Ant’s whole-of-client engagement and thus its moat against competition and its strategic flexibility in the face of regulatory risks. They allow Ant to create at scale and with radically better economics the bundle of information that it used to take a vast, full-service branch network to assemble.
Payments risks being a stranded asset
When Christine Lagarde said last month that the ECB would start work on a “digital euro”, one of her concerns was that, to paraphrase, it should not bankrupt private payment providers.
Whether in China, India, the US or the EU, or many other markets, within a decade the viability of domestic payments businesses will rest entirely on central banks not flipping a switch — like coal plants and oil fields retaining value only so long as their externalities are not priced.
It may be that, like Lagarde today, future central bankers will forever extend their mercies to the payment industry. Some payment companies will say they provide additional value in such and such a way, but the persuasive half-life of those arguments is likely to be short. A well-designed wallet will provide an effective channel for other services, but payments data will be open, and fees trend to zero.
Consumer credit has little to no moat
The consumer credit business model is almost perfectly built for a GPT3-produced pitch deck. Fees and interest start the day a user is acquired, but the costs lie in the future — a couple of raises in the future, by which time an expanding base flatters ratios. All you need is a channel, such as checkout integration or a virtual credit card. If you can work the acquisition cost / early value numbers, you may not even need a proprietary channel.
So there will be waves and waves of new entrants. They will scale fast, since if you spend a lot of money selling money to consumers you will tend to succeed. Switching costs are near zero, since one loan can always pay off the other, and may be negative (the negative switching costs on credit cards, i.e., loyalty points, are currently keeping the airline industry alive).
Some will claim they have “better ML models”, but that’s no longer credible in 2020 when predictive models of this type have become commodities. Others will say they have “better data”, but open banking APIs have proliferated, the Plaid-for-credit-bureaus is here, and the idea that payment channel data will not be subject to open API rules is, again, a slender regulatory reed.
SME + long-term credit needs the whole user
Credit to merchants and long-term or secured credit to consumers is harder to enter than consumer credit. Unit volumes are lower, acquisition more expensive, the risks lumpy and frequent. But pricing the credit right is notoriously hard. Good models are table stakes, but good data is likely to matter more — not just more data, but data with the right features.
Consumer credit and payments data is plentiful, but narrow. It only covers flows, not stocks, and relates to only a part of a person’s balance sheet. A few years ago people spoke a lot about metadata, like phone models or other app installs or social media connection numbers, but even if such proxies do work they are either being locked away or are equal access.
Merchant credit has access to richer data, especially when provided by a merchant platform. That is one of Ant’s core strengths, shared by Meituan, among others — Shopify is making its financial offers ever more prominent. As an old branch manager would tell you though, in evaluating a small business for credit, it helps more than a little to know about its owners as people, above all what resources they can tap in a crisis and how they respond to shocks.
Savings & insurance engage the whole user
Savings (including investments) and insurance build data on users’ entire balance sheet. They build data on the full range of behavioural patterns, including risk appetite and reaction to adverse shocks. The return profile of these services— especially savings and investment — is weighted heavily to the long-term future, beyond the time horizon of venture capital, and they are subject to the kind of early knocks that make fundraising very hard.
These two verticals therefore have a chance at building real moats. In combination with the other services they also unlock a rich space of strategic options and value creation. Most immediate is integrating asset management into a payments wallet — Ant’s first major move, as well as Square/Cash’s and PayTM’s. Another is to extend personal credit or merchant credit against part of an asset portfolio (used by Lufax to recover from P2P).
That strategic option space extends even further if you can add a user’s social graph. Then, for example, an SME loan granting algorithm can be fed not just what a merchant sold or did not, but the assets and financial behaviour of the people close to them, and who they might be able to draw on in an emergency. WeChat Pay forced Ant to build those social graphs half a decade ago. Tencent, though, has been very late to full balance sheet engagement, only launching asset management this year. WeChat Pay now matches AliPay, but WeBank is a distant second to Ant (one fifth the size + profitability).
Savings and insurance are of course far from immune to regulatory risk. But the life of any financial services business is dominated by such risk. Where would you rather be when regulatory risk arrives — narrowly dependent on payments and consumer credit, or with a full bundle of information and the ability to target a full spectrum of services to the whole of users’ balance sheet needs?
Conclusion
Payments and credit have wonderful return profiles in the short- to medium-term, and they appear to be the story of Ant’s IPO. But they generate little to no unique information in themselves, and are at risk of obsolescence (payments) and commoditization (credit), even as operated by Ant. In contrast, savings and investment and insurance produce more limited immediate profits, but a wealth of information, including unique and hard to obtain information.
That information is the reason Ant has its scale and strategic freedom. Its credit and payment businesses will inevitably receive severe regulatory shocks, but when that happens Ant will be able to draw on its engagement of the full balance sheet to find and add new services. WeBank and Synchrony and Brex and others like them have no such options.
Full balance sheet engagement is what the old full service bank branch used to do (with social knowledge too). In the West, the last fifty years saw the unbundling of the information that used to reside in the branch. Many a banking app has tried to bundle the services of the old branch, but Ant has reassembled the bundle of information, without the branch costs, at the scale of a market of 1.4 billion people. Fintech has been promising such a revolution for a long time now. Ant has reached it. Who will be next?
