Alex Hungate
Analyst · Barclays. Your line is now open
Thanks, Jiong. That's great. Let me start by talking about financial services. Just explain a little bit about the lending in particular, which is the focus of your question. So you could see the strong loan dispersals in the quarter, 38% year-on-year, 13% quarter-on-quarter. So we're now at $565 million in the third quarter. So if you approximately annualize that $2.2 billion. So the pace of the lending is going up, and that's because we are layering now the lending through the banks on top of the existing GFin business, where the lending already had a healthy underlying growth rate. Let me start with the GFin business to understand, so you can understand that a bit better because that's a long-standing business. So primarily lending to partners and users across all of the markets, so it's ecosystem based where we are benefiting from the deep insights we have into the user behavior on the ecosystem with a very sophisticated, large language model-based lending model, which are now in just something like 120 different variables from the ecosystem lending decisions. So it's a very different, much more multidimensional type of database than traditional banks would use. That allows us to lend to people like drivers and other gig workers who traditionally have not been well served by banks. So many of them are either unbanked or underbanked because they don't have traditional pay slips. And that's very much in line with our mission as an organization to support them. The penetration of the driver lending by GFin is good, but the models are improving all the time, so we can continue to lend on a risk-adjusted basis. And the risk-adjusted returns overall from GFin are comfortably above the cost of capital for the group. The penetration of lending to merchants is relatively low at the moment. So that's an area where we have a lot of upside. And when Peter answers the second part of your question, you'll see that we have more and more services that are targeted at the merchant and helping them being successful. And the good thing about us as a data science company is that as we provide more and more services to merchants that gives us more and more variables for our merchant lending model, which helps us to – with the same risk appetite start to lend to more and more merchants across all of our countries. So that's the GFin lending model performing well, growing well and producing good risk-adjusted returns. The recent addition on top of that, which is why you're seeing the acceleration of lending is that we're also be able to lend through the banks. The banks are also focused on people who are underbanked or unbanked. And that population is very large in Southeast Asia by some external studies, we estimate like two-thirds of the population of Southeast Asia are either unbanked or unbanked or underbanked. And because we – again, we have very low cost of distribution to them, they are existing users of Grab and therefore, they have an affinity with the brand and because we see a lot of information from them, for example, what they're doing on pay later, whether they're commuting to work, where they live, all these kinds of things we can impute from the behaviors that we see we are able to lend to them. The first lending product, which is now live in all markets, including Malaysia as of this month, November, is the FlexiLoan product, distributed very cost effectively. And then with flexible repayment terms from the point of view of the consumer, it has an extremely high NPS, something like 65. So most banks are hovering at around zero for NPS. So a product with 65 NPS is quite unusual for banking. And so that makes us very successful in terms of the uptake. We are very carefully whitelisting users for that product using the data models. And that means that we will maintain a very careful look in terms of risk appetite. But overall, between GFin and the banks, our nonperforming loan ratio is still around 2%. So that's extremely good performance from those credit models. Then on – obviously, to fund that, the benefit of having a banking license is we get access to very low-cost deposits. And what we've been delighted with since the launch of all three banks is our ability to attract those deposits. We've been able to bring in deposits much faster and with less cost than we had even planned. So you've seen that the positive momentum across GXS Bank in Singapore and GXBank in Malaysia, where customer deposits have grown 50% quarter-on-quarter. So over $1 billion now in the third quarter. And then Superbank, which is the Indonesian bank, which only just launched in July, hit 2 million accounts by October by last month. So a really tremendous uptake. So we are able to raise deposits at a very low cost. And therefore, because of the focus we have on data science and the lending model, you can see very clearly that we can continue this very positive data flywheel for lending. Peter, do you want to...