NU 5.12.26

Do you have a specific take on the main thesis around growth for Nu on your end?

Yep, that makes sense, and it's broadly the same for us. I think on salary, there's this big element of portability, and the government made it way easier. Historically, portability was just not a market a fintech could win in, because the way it worked meant the bank usually had to make direct, one-by-one integrations with large companies. Now that's been really simplified.

I think David Vélez recently commented that there are still some challenges around portability, especially for private payroll loans, and around asset quality. But this is very big. There are two parts to the opportunity you laid out here.

First is just expanding the asset base and the credit book, obviously with the public payroll loans offering lower NIM and, broadly, lower risk.

Second, I think there's a knock-on effect. If Nu actually gets salary portability, that's really going to drive Nu as a primary relationship for many customers. Also, I think it has the potential to lower Nu's cost of capital. If you think about it, if you have your salary being deposited into Nu, the way Nu's deposits usually work is they don't pay interest if you hold funds for less than 30 days or a month. Then, when you hold for more than that, you get a prorated reward on your deposit, which basically matches whatever the rates are, usually 100% of CDI in some cases.

So if you think about it, it's one thing if people use Nu like they use Kaspi, for example, in Kazakhstan, which is mostly for one-time shopping payments, where you'd have a primary bank and then Kaspi would basically be like a wallet for your full salary. So it's very one-off purchases. Versus if you're able to actually have your salary deposited. Obviously, for lower-income consumers, that salary is usually going to be spent throughout the month. So that basically creates a float mechanism for Nu, allowing them to develop a much larger pool of capital on which they might not have to pay rates. The customer might get the salary deposited and use it within the month, and obviously, that comes with zero interest.

Yes and no, I guess. It's very high in Kazakhstan, for example, as well, just knowing about Kaspi. It’s something we worry about. There was a Bloomberg article recently talking about higher default rates happening in Brazil and stuff like that. Talking specifically about that data, I think it's more fintech-driven. It's by virtue of more customers coming into the financial ecosystem broadly, and then obviously there are some bad apples and defaults, and that moves the average up on the rates.

It's tough because real rates are so high in Brazil. But if you think about Nu, I think it's inevitable that those rates are generally way higher than in developed countries, just because they’re very small, one-time, liquidity-driven microloans, where you're almost always going to have a high effective yield and interest rate on the product. But in absolute dollar terms, it's broadly not as high.

I don't know what the average size is for Nu, but thinking of MercadoLibre (MELI), I think MELI reported about $400 as the average exposure for their consumer loan customers. Obviously, if you have a high rate on that, it might be sub-$100, which would still be high, right? But in absolute terms, it's payable. It would be another thing if customers were taking out $5,000 in loans carrying exorbitant interest rates.

I think the other element is the increasingly high risk of regulation. I might be mistaken, maybe you can correct me, but I think it was Brazil that capped interest rates.

Oh, wow. And then obviously the regulation with the FGTS, where they basically restricted that market as well, and origination fell by, I think, 60% instantly on those loans. So stuff like that, I’d be worried about.

Broadly, I think it's OK. They're actually fairly conservative in how they manage their balance sheet. Their LDRs, I think, are in the 40%-50% range if you only look at the interest-earning portion. A large part is also Nu, having $2 billion, it might be more now, I don't remember, but a couple of billion in just cash sitting on top of the required regulatory capital. I think even under a fairly heavy macro downturn, Nu probably does OK, unless obviously it's something really horrible like an '08 crisis.

Yeah, also on regulation, there's another element that almost mirrors what regulators in Kazakhstan did with their banks, where Brazil increased the tax rate. I think it's 30-something percent right now, and then they increased it to 40% and eventually 45%. So that's another element. Maybe because Brazil is actually very forward and progressive when it comes to regulation, Pix is an example of this, but at the same time, they've really driven this fintech market and financial inclusion. Maybe now it's a pivot to actually try to bring more fiscal discipline, because, broadly, if you look at the macro, Brazil isn't the most fiscally disciplined country, I'd say, in terms of government debt and stuff like that.

Yeah, I agree with that. We're long MELI as well, but on the credit team, I do agree. You have Osvaldo Giménez from MELI as the head of credit or head of fintech, and he doesn't really have a banking background, which is OK. I mean, the CEO of Kaspi never really had a banking background, either, but I do think the team broadly is probably weaker.

Management did comment that, according to their data, they now have the highest deposits across all financial institutions. I don't know how true that is; maybe there's a different methodology or something. On products, you're right that Nu does have a broader range, though on a recent call, MELI's management said they were planning to launch payroll loans.

The other element, I think, is branding broadly. Nu invests a whole bunch into branding, and it’s a proper fintech banking brand. Whereas for MELI, Mercado Pago was always a thing, and Pago is probably for slightly lower-income customers, even more so than Nu.

One interesting thing about the data advantage is that, historically, MELI had a significant data advantage over Nu. But Nu still generally did better, just because they were better at using data than MELI. Nu would have worse raw data, but their ability to parse it and leverage transformer AI models was generally better. Recently, I've read some expert calls in which former MELI and Nu employees discussed MELI stepping up its game by improving data parsing and using its rich dataset more efficiently.

But then, with Nu, its ability to integrate with Amazon (AMZN) and, recently, with Shopee is probably an incremental advantage. Finally, the fact that Nu is much more likely to become a customer's primary bank than Mercado Pago probably has knock-on effects on which data matter in credit underwriting. If Nu does end up becoming a true primary bank for the majority of its customers, you might argue that even though MELI has MercadoLibre as a dataset that Nu doesn't, by virtue of Nu being a primary bank, it might have an actual advantage in terms of more relevant data, if that makes sense.

Yep, that makes sense. Obviously, that's a big advantage for SMEs and merchants on the data side, but also for cross-selling and broad access to that relationship by virtue of MercadoLibre. But I generally agree.

I think the most valuable data is actually off-platform for MELI, and Nu obviously has it through credit card spending as well. So here you can probably argue that Nu does have an actual advantage.

I'm curious what you think about the ARPAC leg of Nu's growth, especially in Brazil, where they have north of 60% of the adult population. It seems, increasingly, in Brazil, the story is about actual revenue per user rather than user account growth. I'm just curious what you're thinking on this as well.