Description

It’s not every day you find a company that is a 20%+ long-term compounder and accelerating Y/Y growth while trading below 3x 2023 sales. What’s even more amazing to me is that despite all the headwinds going into 2022, they will still be profitable. I present you STNE, the #1 payment company in Brazil. 

STNE was written by Crow back in 3/25/2020 and I recommend you reading his write-up if interested in the name.

Thesis Summary: Stone is a Brazilian payment acquirer that offers the lowest payment fees while offering the best customer services and best software for SMB’s. The company strategy is to make their profits from financial services such as prepayment of card receivables and working capital loans. Recent credit issues combined with poor mgmt. communication has set the wrong market expectations, causing a big surprise 3Q21 results with significantly lower-than-expected margins, despite posting record-breaking top-line numbers. The company lost over $15 billion in market value since those issues arose (down a total ~80% YTD), and its trading at its historical lowest sales multiple while revenue growth is accelerating. With management’s creditability and sales multiples at an all-time low, this has created one of the best LONG opportunities we see today to buy the Square of Brazil for only 5-8x Fy23 EBITDA. We think that by 2023 those issues will be well past, and the company will be back at 40% EBITDA margins and generate R$ 5.5 billion of EBITDA (before financial expenses) with their Credit product back in full force, and the market will quickly get excited again. The stock should re-rate back to ~10x sales (a 30% discount to its ~15x sales multiple in 2019) or a 20x EBITDA multiple, which is in-line w/ 2019 levels and well below the 40x multiple of last year when the market was excited on the Credit opportunity. Founders of Stone still own 17% of the company and the last time that management sold shares was in June when the stock was above $70/share.  

STNE price chart

Company overview: STNE currently has 10% share of Brazil’s total retail card transactions, and 50% of all e-Comm transactions go through Stone. The company makes money charging an avg of ~60bps on every payment transaction and makes another ~20bps from POS device rentals that merchants need to accept card payments. Stone predominantly makes their profits from prepayment of credit card receivables, that’s about another 100bps per transaction, and since 4Q19, they’ve added another product called “Credit” which are working capital loans. That would add another 40-100bps per transaction. In total Stone has a ~280bps of net take rate potential on every transaction. (I’m also excluding digital banking solutions, their issuing bank card and business insurance since those are fairly new).

Stone’s strategy has always been to offer low payment fees with best customer service and provide more software tools to its clients. They’ve innovated with what they call Hub distribution model, which are hyper-local distributed teams that are located very close to their merchant customers. This strategy allowed them to react very quickly to competitive dynamics, and to gain speed to acquire and service new customers. For example, before this model… incumbents would sign up a new merchant and the merchant would only receive their POS ~15+ days later. Incumbents would outsource the POS device delivery and sometimes even sales agents, and they all outsource their call center for customer services. By being vertically integrated (except POS manufacturing), the company is able to attain an NPS that is 2x higher than is nearest peer.

The company has a vast portfolio of smaller SaaS products they’ve either built or acquired over the years on the software side. Because Stone was growing payments so fast, they’ve never focused on integrating all those SaaS companies into a single solution, so they ended up cross-selling to merchants as separate solutions. Important to understand is that despite Stone not yet offering an integrated solution like Square does in the US, all of Stone’s peers and incumbents don’t have any software solution whatsoever. The only thing they offer is a simple financial dashboard. So despite not ideally integrating all-in-one, Stone still has a large software moat. They’ve recently acquired Linx, the largest retail ERP software with 45% retail share, to strengthen their software moat. As of 3Q21 and including Linx’s ARR, Stone has a total ARR of ~$230m (21% of total sales) and growing 27% YoY. The company is currently working on customizing Linx’s solutions with Stone’s existing portfolio to finally bring a single integrated all-in-one solution to SMB’s, something that does not exist in Brazil.

What created the opportunity?

1)   The company messed up their new working capital loan product badly, which now has NPL’s standing at 48%.

2)   Stone also surprised the market when they reported 3Q21 numbers with higher-than-expected opex.

3)   Lastly, Stone reported much higher than expected financial expenses (this is their funding cost for their prepayment product).

Those issues have led management to lose a lot of credibility, and I admit, they deserve to lose credibility. But the extent of the market losses far outpaces a reasonable price compression due to their missteps. To illustrate this into actual numbers, their Credit NPL’s equates to ~$150m and even at 100% default, that would be $300m. This compares to STNE losing over $15 billion in market value since those issues arise. Let’s address each of the 3 items above.

What was the issue with Stone’s Credit Product?

• In 4Q19 they started offering working capital loans in addition to their prepayment of card receivable’s product. These loans were collateralized by future credit card receivables. When you make such loans, you’d effectively “lock” 100% of the merchant’s future receivables and they would go to a specific bank account where the lender controls.

• During Covid, the company saw an opportunity to grow this product very aggressively. Still, they did a poor job underwriting the risks, and later this was extrapolated by the 2nd wave of Covid, which caused many merchants (and some peers) to take advantage of a loophole in the system (aka “collateral leakage”).

• The old regulation allowed for “bi-lateral agreements” between merchants and lenders instead of registering those loans at one of the public registry databases where everyone in the industry could see. When the 2nd Covid wave hit Brazil, merchants desperately needed more capital, but they couldn’t get another loan since all their receivables were locked, regardless of the previous loan amount they received. Merchants then switched acquirers to avoid Stone’s controlling their cash flows, and some acquirers purposedly took advantage by not “checking” the public registries of receivables to see if such merchant already had receivables locked with another provider, because the previous system allowed for “bi-lateral” agreements. So the regulation wasn’t evident on whether you had to check the registries or not.

•  On June 7, 2021, the new regulatory framework started, and it was supposed to fix this loophole by forcing all previous “bi-lateral” agreements to be registered at one of the 3 public registry databases, to be then followed by a waterfall order of priority (first loan comes first). Stone knew they would get priority on their loans since they were first lenders.

• However, these 3 registries were not operationally capable of handling the number of inflow contracts that came in on June 7, becoming a mess in the first few weeks.Stone temporarily lost priority on those receivables, and because they are known for its superior customer service, the company didn’t want to seek legal action against its customers, afraid of a PR backlash disaster. Stone saw itself in a challenging situation of recovering these loans in the short term and had no choice but to write them as 60 days NPLs, effectively 48% of its loan book in 3Q21.

The new regulatory framework does 3 main things that are positive for the industry and for when Stone re-start offering Credit:

1)   Removes bi-lateral agreements, forcing new loans to be registered and checking if receivables are already locked.

2)   Instead of locking 100% of the merchant’s receivables, it will lock only the “loan amount”. This was a key issue that made merchants switch away from Stone’s acquiring services.

3)   Will also allow for future debit card volumes to be used as collateral.

These changes will not only fix the collateral leakage issue, but it will also significantly increase the TAM opportunity for Stone’s working capital loans.

Why should we believe Stone will be a good Credit provider if they’ve messed up so bad in their first trial?

•  It’s a fair criticism, and indeed the company has a significant degree of fault here. However, high-growth companies have a high trial and error level. As problematic as this sounds (48% of the loan book at risk of default), the actual total loss so far accounts for less than $150m. Even at 100% default, the total loss would be ~$300m. This compares to the company losing over $15 billion in market value since this issue arose.

• We believe Stone can re-do this product the right way, and they recently purchased a credit startup Gyra+ which seems to have a very good technology that is specialized in working capital loans.

• Linx acquisition is also a huge plus for improving the underwriting of this product. Linx has 45% of the retail ERP market and access to deep merchant’s data at every retail market segment.

•  STNE will start to test disbursement of loans again in 4Q21, and we think by 2H 2022, this product will be at scale again.

• By 2023, we think the company will disburse at least 1.2% of TPV in working capital loans (Square does ~2% in the US).

Higher than expected opex in Q3 – is Stone a structurally lower margin business?

We think Stone is structurally the same business in terms of margins, but the integration of Linx and growth opex are masking the underlying true business margin. The higher funding cost is a temporary issue, and rates will eventually decrease, so their spread will increase again. The funding cost, which came at 22% of revenues vs 15% previous quarter, was the main concern in investor’s minds because the interest rates in Brazil were only at ~5% during Q3, and we know it’s going to 10-12% in early 2022.

However, what is remarkable is that despite this considerable headwind coming into 2022, the company will still make plenty of money from the prepayment product. Stone has a cost of capital ~2-3% above Brazil’s CDI rate. Let’s assume the 12% high-end cons 2022 and say their cost of capital will be ~15%. STNE typically prepays ~40-50% of their Credit card volumes (or ~30% of total TPV).

If we assume 57 in turnover days (same as disclosed by Cielo during Q3) we arrive at a 20% gross yield (before cost of capital) based on 3% fees we think they charged customers on avg during Q3, where Brazil int rate was ~5.2%. Stone doesn’t disclose prepayment pricing for the SMB, but we can use how they are repricing the TON machines for micro-merchant as a data point. Using Giga TON pricing increase Nov vs Oct (Giga TON is also vastly used by many SMB customers), we know they’ve increased prepayment fees by 30%.

Assuming this 30% increase across the board, we would get a 3.9% avg fee for 2022 vs 3% in Q321. At 57 turnover days, that implies a 27% gross spread. Subtracting the 15% cost of capital, we arrive at a net spread of ~12% which is flat w/ my estimates of 12% in Q3. The acquirers are not repricing their prepayment fees based on next-month rate, they are looking at the full-year curve, so I think this is a fair estimate for 2022. Regardless, even if I’m wrong, I think focusing on the 2022 spread is the wrong thing. At some point, probably during 2H 2022, interest rates should start to decline, driving the cost of capital lower, while Stone will be able to keep new higher repriced fees, which at that point, the headwind will become a tailwind. We think this will take place before we go into 2023.

How do we answer one of the main questions investors have. Is Stone a structurally lower-margin business?

We didn’t get any color on Linx numbers other than they are negatively impacting margins. If we make some reasonable assumptions for what was Linx’s opex during Q3, then remove some one-offs, and remove the R$ 130m growth opex STNE disclosed, we conclude that Stone is fundamentally the same business and the market is wrongly assuming they are structurally a lower margin business going forward.  

See below that after some adjustments to 3Q21 numbers, STNE would have a 47.5% ebitda margins, in-line w/ 3Q19 48.5% ebitda margins (both after financial expenses). 3Q19 is comparable to 3Q21 since both quarters didn’t have any Credit revenues.

We think investors got too accustomed to Stone’s high EBITDA margins and are forgetting the bigger picture: This is a growth company in a rapidly growing market. Card payments in Brazil are growing very fast Y/Y -- 25% in 2021, 20% in 2022e, and expected DD for 3-5+ years. Even during 2015/2016 when GDP contracted MSD and retail sales declined mid-to-high digits, card payment volumes grew 9% in 2015 and 15% in 2016. The R$ 130m in growth opex during Q3 equated to 10% of Stone (ex-linx) total sales. But those investments should be seen as market share gain acceleration, not a margin headwind. 10% of sales in opex investments can easily be offset by operating leverage on a company that is going to CAGR 45% sales during 2021-2025.

One of the key investments Stone did was in its in-house call center, to make a commitment to answer customer calls within 5s. We think this was a key reason why Stone had a record net adds in Q3 for both segments (80k new SMB merchants and 215k new micro-merchants), interestingly at the same time that PAGS had its lowest net adds ever. STNE currently have 1.38m active merchants (846k are SMB-merchants and 542k micro-merchants) with 50% of its SMB base (or 30% of all active merchants) are using Stone’s new digital banking and 75% of those are settling payments in those accounts. STNE doesn’t have a banking license, so they can’t use these account balances to fund prepayments, but that could also be an interesting future source of capital to lower the cost of funds, but this also proves that STNE can over time own more of their customer's financial ecosystem.  

Some other items worth mentioning

The recent PIX payment layer introduced by the central bank is another miss-understood topic. We view as a major tailwind for the payment industry, which is helping accelerate introducing end-consumers that were predominantly cash-only into electronic payments, weather be via cards or a digital wallet using PIX/QR-Code.

The OpenBanking initiative starting in 2022 is very bullish for fintech companies like Stone. It forces all the banks to open up their customer’s data via APIs available to any fintech player, as long as the customer gives the consent. This will allow Stone and others to access consumers’ and merchants’ financials that were previously siloed and only available to banks, to provide better product offerings.

One argument we hear is that competition is much higher. That is wrong. The competition is increasing indeed, but it’s mostly on the long-tail segment, which affects mostly PAGS, as they are almost entirely micro-merchants. Can PAGS compete with SMBs? Not really. PAGS lacks software (have none) so they can only compete in pricing on the SMB segment. At the end of the day, this is still a Cielo and Rede story. These 2 still have > 50% of the market, and they too lack any fundamental value proposition other than pricing. These 2 peers are controlled and run by bank executives who lack technological know-how and can’t attract engineering talent.

VALUATION

Assumptions in the model.

• We expect working capital loan volumes to reach 1.3% of TPV in 2023, which should be reasonable given SQ’s currently ~2% and Brazil merchants need more of those loans due to the installment market dynamics.

• We also think our opex margins are reasonable and it does include growth opex to continue and are not accounting for Linx synergies, which would be further upside.  In YE2020 Linx had 4k employees, we think there is a lot of fat that can be cut there.

• We expect Brazil’s interest rates to be down to ~7.5% 2023 vs 10-12% in 2022, thus lowering funding costs.

It’s important to note that Bloomberg tracks EBITDA multiples (excluding financial expenses). So when you look at historical ebitda multiples on Bloomberg, those exclude financial expenses (e.g., in 2020, STNE reported EBTIDA margins were 54% excluding financial expenses or 46% if adding them). If you want to use EBITDA after financial expenses for valuation purposes, you must remove company debt, since it is non-recourse. 

At the current price of $15/share, you get to buy STNE at only 5x 2023 EBTIDA before or 7.5x after financial expenses. That’s one of the best bargains we ever found for a company growing top-line at 40% CAGR over the next 4 years. At 10x Fy23 sales, we are valuing STNE at 12x gross profit, getting to a stock price target of $63 and $84 for Fy23 and Fy24 respectively. The ~12x gross profit multiple is in-line w/ SQ and GPN if you take into consideration that STNE is will be growing faster than both. (Gross profit is typically a good metric to compare different payment companies, due to differences in revenue recognition).

Key risks: 

- Inflation .

- market volatility due to 2022 presidential elections.

Catalyst

- Brazil CDI rates start falling sometime in the second half of 2022, after peaking in the first half.

- Company continues to gain market share and post positive business fundamental metrics that matter (not short-term net margins). 

- Company shows progress on Linx acquisition (weather on cost or product synergies).