Description

Thesis: Dlocal (DLO) is a long because the primary claims made in the recent Muddy Waters (MW) short reports, which have driven a roughly 40% decline in the stock price, are verifiably false (MW reports are available here). This write-up focuses on the MW reports, for more background on the business please see the VIC writeup by alum88 in January of 2022 which generally aligns with our views. If the MW claims are incorrect, DLO is a high ROIC business (low invested capital with >35% EBITDA margins) with sustainably high growth rates (historically >50%, prospectively we believe >35% for at least the next few years) that will have > $2 per share in net cash by YE23 available at ~20x 2023 earnings. We have great respect for Muddy Waters and have followed their research for over a decade, so we don’t post this lightly. We also note some places where we do not have clarity or we hope the company makes increased disclosures.

DLO Overview: DLO is a payment service provider (PSP) that offers merchants payment processing, FX management, fund collection, fund settlement, fund disbursement, and additional features to minimize friction for merchants, increase conversion rates and reduce fraud through a single API. DLO specializes on cross-border payments in emerging markets, which is a complex and (as a result) lucrative sub-segment of the payment industry. DLO enables global enterprise merchants to get paid (pay-in, traditional ecommerce) and to make payments (pay-out, payments to vendors such as Uber paying drivers) within emerging market countries. The current split is approximately 1/3^rd local-to-local payments and 2/3^rd cross-border payments, 3/4^th pay-in (ecommerce) and 1/4^th pay-out (dispersions to partners). DLO has ~600 clients with a focus on large enterprise customers. Top customers include technology players such as NFLX, MSFT, META, AMZN, SPOT, GOOG, UBER and ABNB. DLO IPO’d in June 2021 at $21/share.

SUMMARY:

The primary claims made in the MW report are as follows (the quotes are directly from the original report and listed in the order in the MW report), along with our primary findings:

1. Conflicting total processing volume (TPV) disclosures: “DLO repeatedly makes contradictory statements about its historical TPVs. We believe that these contradictions exist because DLO is misstating its financials and TPV, and that the web of lies has become too complicated to be repeated consistently.”
   1. DLO’s total TPV each year in every filing has been the same. The variation in new merchant TPV by year is because MW is comparing two different ways to breakdown TPV. The initial way used in the Draft Registration Statement (DRS) is not a cohort analysis but a waterfall analysis. New merchant TPV in the DRS includes (1) TPV from merchants in the current cohort year and (2) TPV from merchants in the prior cohort year when that merchant has been a customer for less than one year. The DRS analysis basically cuts TPV to show “same-merchant TPV growth” (merchants over one year old).
   2. It’s also worth noting the DRS filed on March 2, 2021, did not include any 2020 financial numbers as the 2020 audit was not finalized by PWC until April 13, 2021.
2. Conflicting Receivable Disclosures: “DLO disclosed two different sets of numbers for its foreign currency receivables. The foreign currency receivable discrepancy is approximately 10% of its total FY 2020 foreign currency receivables. As we detail infra, we believe that FX is the primary tool with which DLO inflates its revenues and profitability.”
   1. We believe that the $4.2M difference in “Net Monetary Position” is due to accounting for a new long-term lease signed in Aug-2020 but that didn’t start until Apr-2021. The leases liability is also $4.2M and IFRS says that a lease liability is considered a monetary liability if denominated in a currency different from the company’s functional currency
   2. We do not consider this to be a material point as the difference does not impact revenue or profitability. At the same time, we would like to have more disclosure on this point from the company but do not consider this to be a material point in the report.
3. Subsidiary Account Conflicts: “DLO has three primary operating subsidiaries – DLocal Ltd (Malta) (“Malta Operating”), DLocal LLP (UK) (“UK LLP”) and DLocal Corp (UK) (“UK Corp”) – which in turn sit at the top of a network of local subsidiaries. In 2019 and 2020, these subsidiaries respectively accounted for ~94.9% and ~90.5% of consolidated revenue. UK LLP’s 2020 financials show that as of December 31, 2020, it owed Malta Operating $15.7 million. Malta Operating’s 2020 financials show that as of December 31, 2020, it is owed by all related parties a total of only $12.4 million. In other words, UK LLP’s financials show that it owed Malta Operating 26.6% more than Malta Operating’s financials show it could be owed.”
   1. MW’s analysis didn’t include data from DLO’s third operating subsidiary, Dlocal Corp UK, which had a receivable with Malta Operating. With that 99.1% of Malta Operating’s 2020 net receivable is explainable.
   2. Furthermore, on 11/21/22 post the release of the MW report, DLO’s Malta Subsidiary filed their 2021 annual report. When taken together, they explain > 99% of the inter-company AR and A/P (they shouldn’t add up to 100% given DLO has many other subsidiaries), which disproves the MW claim.
4. Let’s Change the Facts – Loan Agreement with Insiders: “DLO altered its accounts to avoid the negative optics of making a large loan to its senior management pre-IPO. The falsehoods have to do with the timing and funding source of a large pre-IPO stock option exercise by the founders.”
   1. The facts were never changed. The disclosure MW points to on 6/4/2021 is the same disclosure that DLO had since its DRS/A on 4/13/2021, when its 2020 financials were first filed.
   2. Looking at all the transactions, the true events seem relatively benign. Singer/Kanovich were issued stock options on 1/31/2020. Later that year when it became clear that DLO was likely going to IPO they took a loan from the Company to exercise those options for tax purposes (standard for a pre-IPO company) but economically and under IFRS it’s considered a stock option until the loan is repaid. The loan was repaid in March and April 2021 (partially before the PrimeiroPay acquisition). The two wire transfers MW showed on December 17, 2020 were not Singer/Kanovich repaying the loans but inter-company transfer from DLO’s Citi account to DLO’s Bank J. Safra Sarasin account. This wire transfer was likely made to make the loans official for tax purposes but no money actually changed hands between DLO and Singer/Kanovich.
   3. Additionally, every contact that have spoken to has had positive things to say about company president Jacobo Singer and CEO Sebastian Kanovich.
5. Why the Lies? Possibly There’s a Bigger One! – Acquisition of PrimeiroPay: “Regardless, it does strike us as an interesting coincidence that Kanovich and Singer funded these payments at the same time DLO sent $38.7 million to purchase the assets of PrimeiroPay.”
   1. PrimeiroPay is a real asset that generated €2.0M of gross profit in 2018 and has good reviews from its customers. PrimeiroPay’s shareholders are not linked to Kanovich or Singer. The acquisition was independent of options being exercised.
6. Skepticism on Take Rates: “DLO’s Take Rate is an outlier among peers – i.e., it seems too good to be true. DLO’s Take Rate grew significantly in the year prior to its IPO – an IPO in which insiders cashed in $325 million of stock – despite conditions in its business implying that its Take Rate should have actually decreased. In 2020, FX gains and fees contributed approximately 290 basis points to DLO’s take rate. We believe that the contribution should have been no more than approximately 150 basis points, and likely much lower. We therefore strongly suspect that DLO’s revenues are materially overstated.”
   1. DLO take rates are in line with other cross border payment providers. The take rate MW calculates for DLO’s primary private competitor Ebanx, only used Singapore financial statements and excluded Ebanx’s Brazilian subsidiary, a glaring error. When consolidated the financial statements between the two subsidiaries (shown below), Ebanx’s take rates are very similar to DLO’s.
   2. Data and quotes that shows that DLO take rates are similar and in some cases below competitors on an apple-to-apples basis. Ebanx’s publicly available pricing sheet for SMB shows that it charges a 5.5% FX fee for pay-ins.
7. Concerns about Client Funds Segregation: “In 2020, DLO’s last year before going public, it paid its pre-IPO shareholders a special dividend of $15 million and made a $31.5 million loan to two shareholders (discussed supra). When running a cash flow reconciliation net of movements of merchant funds-related accounts, DLO’s cash flows do not reconcile. We calculate a $3.3 million deficit in DLO’s ability to fund its dividend. This problem is mirrored at the Malta Operating subsidiary level, at which we calculate a $4.1 million deficit in the subsidiary’s ability to fund its cash uses.” MW’s second DLO report, released on 12/1/2022, focuses on this issue as well.
   1. DLO paid its $15M special dividend to pre-IPO shareholder during Q3’20, likely >4.5 months before MW’s “so-called $3.3M deficit”. At the end of Q4’20, DLO also had $52M in cash and financial assets.
   2. DLO’s own financials imply that merchant client funds settle in about a week, consistent with comments from customers, formers, and competitors. Thus, it seems highly unlikely that this cash could be used to pay a dividend without DLO’s group of highly sophisticated customers noticing it.
   3. MW’s analysis of trade’s receivables vs. trade payables doesn’t imply a cash deficit in merchant funds. All it shows is DLO had a positive change in net trade working capital in 2020 (+$3.4M). In 2021, DLO had a much larger negative change in net trade working capital (-$19.4M) but that doesn’t mean that merchants are “stealing cash” from DLO. Many other PSPs such as Adyen and Nuvei had net trade working capital cash flow benefits in 2020.
8. Governance Deficiencies: “DLO strikes us as being unduly complex, having poor controls including an over-reliance on manual processes, deficiently governed, and a regulatory arbitrageur.”
   
   1. Lots of small things here that management can further clarify that is not material. DLO will likely switch its UK auditors to PwC, DLO will likely be regulated in the future by the FCA since the UK is no longer part of the EU after Brexit and fixed its Malta Tier 1 requirement, which DLO admitted in its 20-F already. DLO's related party revenue is zero YTD in 2022 (down from 2% and 1% of revenue in 2020 and 2021, respectively) and is separate from Directa24/AstroPay. Today, AstroPay is an e-wallet or alternative payment method (much like NETELLER or Skrill) as such there is nothing inherently wrong with DLO processing this type of transaction for its merchants.
9. Insider Selling and Insiders Fleeing: “The IPO and secondary were intended to enrich insiders…since June of this year, there has been an exodus of senior executives”
   1. Management sold in the IPO and subsequent secondary, part of which was likely to repay the pre-IPO company loan used to exercise their options. We would note that outside of the $21 IPO, we do not believe a single share has been sold by insiders at a price that is not at least 2x the current price for DLO.
   2. We don’t begrudge General Atlantic (GA) selling stock at $52.25 per share, in what was surely a significant windfall for them. GA also is incentivized to sell to lock in its significant profits to its investors. GA also was the primary investor in Adyen (ADYEN NA) and based on filings GA fully exited the position by 3/31/2020, less than two years after Adyen’s IPO (6/13/18), even though Adyen’s stock price has continued to climb.  
   3. We acknowledge that DLO has had management turnover, although most of the top management has been consistent. At the same time, our calls with former employees indicate that at least some of the turnover, such as the former COO, was driven by the company and not the individuals.

FURTHER DETAILS ON KEY POINTS:

The most significant points raised in the report in the report was the subsidiary account conflicts, the skepticism on take rates, concerns about client fund segregation (which was confirmed by the follow up report on 12/1/22), the exercise of options, and the PrimeiroPay acquisition. If those issues are disproven, many of additional concerns (insider selling, governance, etc) carry significantly less weight. As a result, we will focus most of the rest of those writeup on those topics but are happy to address any other lingering concerns within the Q&A.

Subsidiary Account Conflicts:

• This claim is provably false.  MW accounting claims are incorrectly framed, and management have already given investors disclosure to debunk the bear case. These points will all carry over when talking about MW’s (incorrect) claims about client funds.
  • The MW report compares two individual fillings (“UK LLP” and “Malta Operating”). UK LLP’s 2020 financials show that as of December 31, 2020, it owed Malta Operating $15.72M. Malta Operating’s 2020 financials show that as of December 31, 2020, it is owed by all related parties a total of $12.36M. Thus, there is a $3.3M delta between these filings.
• The problem with MW’s analysis is it didn’t include the third operating subsidiary, Dlocal Corp UK (“UK Corp”).
  • UK Corp’s 2020 financials show that as of December 31, 2020, UK Corp is owed $3.48M from Malta Operating (Dlocal Limited).
  • So, Malta Operating is owed $15.72M from UK LLP and owes $3.48M to UK Corp, so netting that out implies Malta’s net receivable (just from those two subsidiaries) would be $12.24M, which is 99.1% of the actual Malta Operating receivable of $12.36M.

Below is Note 13 from 2020 Dlocal Corp UK:

• Dlocal released the 2021 annual report for its Malta Operating subsidiary on 11/21/2022 after MW’s short report was released.
  • In that report, Malta Operating shows the gross receivables from related companies and the gross payables from related companies rather than the net receivable from related companies in the 2020 report.
  • From this NEW DISCLOSURE you can see that inter-company transactions even easier.
    • 99.2% of Malta Operating Gross Receivable due from related companies in 2020 is from UK LLP
    • 99.8% of Malta Operating Gross Payable due from related companies is from UK LLP
    • Exactly matches the 99.1% net receivable percentage from above.

• Using the three biggest subsidiaries explains >99% of the inter-company A/R and A/P, which makes sense given DLO’s 20+ other subsidiaries (it would be more suspect if only those three were exactly 100%).
  • The core points are the (1) the MW report has some obvious accounting errors and (2) management can and already has given some more disclosure to disprove the bear case.

Below is Note 9 and 12 from 2021 Malta Operating

Skepticism of Take Rates – Dlocal vs. Ebanx:

• MW’s calculation of Ebanx’s take rate is incorrect          
  • MW Report: “We consider Ebanx to be DLO’s closest comp because a) until 2020, it primarily focused on cross-border payments from / to Brazil… Assuming that Ebanx’s 2019 TPV was $2 billion, and that 90% of it was cross-border (which we think is DLO-favorable for this comparison), Ebanx’s FX revenue, which is reported on a gross basis, represented ~0.60% of the hypothetical cross-border TPV. (That rate drops to 0.55% if assuming all $2 billion was cross-border.)”

• MW use Ebanx’s private Singapore financial statements (EBANX Pte. Ltd.) to show that Ebanx had revenue of $70.6M and Gross Profit of $12.0M in 2019. But these Singapore financial statements are not fully consolidated financial statements.

• From Ebanx’s Singapore financials we can see that $49.3M and $41.3M of costs in 2019 and 2018, respectively, are to related parties, which are to Ebanx’s Brazilian subsidiary.

• MW completely excluding Brazilian (EBANX S.A.)  revenue and costs from its analysis. From Ebanx’s Brazilian private financial statement in 2018 (2019 unavailable) we can see that Ebanx generated R$303.965M in revenue ($82.6M) and R$168.847M of gross profit ($45.9M) from this subsidiary in 2018 and even more in 2019.

• Can also see that the Brazilian subsidiary has inter-company revenue from its Singapore subsidiary (EBANX Pte. Ltd.) for R$151.854M, which at 2018 average BRL/USD exchange rate is the $41.275M of inter-company cost we see in the Singapore financials.
  • Note: Ebanx implied BRL/USD exchange rate is 3.679 vs. simple daily average of 3.654. The 0.67% difference is easily explainable given BRL depreciated throughout the year and revenue likely more weighted to H2.

• Combining the Brazilian and Singapore financial data as well as numerous public articles that quote Ebanx’s 2018 TPV at $1.5Bn and 2019 TPV at $2.2Bn we can get Ebanx’s actual gross and net take rates for 2018 and a good approximation for 2019.

• Articles imply that in 2021, Ebanx processed ~$7Bn in TPV (higher than DLO) with revenues $239M to $288M (similar to DLO).
  • Thus, Ebanx’s gross take rate in 2021 was likely between 3.4% and 4.0% compared to DLO’s at 4.04% in 2021 (<65bp difference).

Ebanx’s Own Take Rates:

• Ebanx Fee Chart for SMBs (<$1.8M in TPV).
  • DLO defines Enterprise merchant as >$6M of TPV
  • Ebanx charges a 5.5% FX fee for Pay-ins and 3.0% of FX Pay-outs
  • DLO's Shopify pricing sheet (for SMB too) is pretty much always cheaper than Ebanx.

Critical of FX Fees:

• FX fees are a big part of DLO’s take rate that has always been known by the market and is normal industry practice.
  • JPM Initiation Report:
    • “In our view, part of the higher take rate for DLocal is due to FX spreads” – JPM Initiation
    • “FX spreads have been actually moving up in the last decade, and spreads for non-export contracts are running close to ~120bps in the country, according to a Central Bank study”
• FX fees are sustainable
  • Ebanx CFO in Private Meeting: “Think FX spread sustainable given complexities.”
    • Complexities are of the game.
    • If Africa becomes Sweden then that is not good for FX spreads and Ebanx. But if LatAm/Africa keeps currency risk/volatility then there is value for FX conversion so the merchant gets money in USD/EUR.
    • They are all across EM all with different regulations and APMs.
    • FX is very regulated and do all that under the hood. Ebanx assumes the regulatory risk and settles the money back in hard currency so it is a powerful value prop.
    • An int’l card used XB would have >100bps of FX spread imbedded in it and they bring more value than that.
• MW also says that Ebanx’s only makes “21% of total revenue from FX fees in 2019” compared to 45% of DLO’s revenue in 2020.
  • MW uses data from Singapore financials which excludes Brazil’s mix of FX revenue (which is unclear).
    • Ebanx CFO in Private Meeting: “1/3 of gross take rate is FX spread”
  • Given DLO’s mix shift to more local-to-local processing today, FX spread probably makes up ~40% of DLO’s gross revenue (~700bp difference vs. Ebanx).
• DLO’s FX fees are likely sustainable due to the Visa/Mastercard large FX fees
  • Gross take rates for cross-border transaction are 200-300bps higher than a domestic transaction. In a cross-border card transaction the networks get an extra 75-125bps and acquirer gets another 75-225bps. 
    • VP of Partnerships at Mastercard: “we isolate ourselves from FX by charging the 75 to 125 basis depending on the country and region. So basically, what happens is when the transaction happens, the merchant is charged, let's say, 2%, 2.5%, 3% depending on who the acquirer is.”.
  • For a cross-border card transaction, Stripe charges an extra 200bps in Mexico and an extra >200bps in Brazil. 

• • Cross-border fees are even greater than 200bps in higher risk and/or Emerging Market countries other than Brazil/Mexico adds
    • For a cross-border card transaction in high-risk countries, Verifone (2Checkout) charges +300bps compared to a +200bp in low-risk countries.
    • For a cross-border card transaction in high-risk EM geographies, PayPal charges an extra 400-450bps for Pay-ins and an extra 300-350bps for Pay-outs (compared to +150bps in the US, +161bps in Brazil).
      • On top of that, PayPal also charges for currency conversions (another 300-400bps).

DLO’s Take Rate Makes Sense:

• Tons of customers quotes that DLO’s gross and net take rates (~4%/~2%) make sense and are in-line with peers.
  • Head of Payments & Director of Product at Customer 1:
    • “At [Company XYZ], I lead the Payments biz unit across Payins, Payouts and Payment platform incl Product, Engineering, Data, Design functions.  We are a current customer of dLocal. We do about $150-$200M transactions via dLocal. That is GMV, not spend on dLocal. Spend [Gross Take Rate] would be 3-4% of that.” – old statement
    • “We do about $300M in transaction volume, and spend about 4% on that, so about $8M [with Dlocal]…[for] “payment processing and cross border payments.” – recent statement
  • Director Of Revenue at Customer 2: “dLocal would be about 10% cheaper than EBANX, and they would be about 15% cheaper than like PPRO or allpago, a little cheaper now also against PayU, probably 5% to 10% cheaper. They are really the cheapest payment provider for high-volume processing in the region overall”
  • Former General Manager of the Hospitality Solutions at Customer 3:
    •  “I think all-in costs was not that different by market. I believe I'm recalling in the range of like 2.5%, maybe 3% to 7%, depending on the market of the transaction.”
      • Q: And I'm assuming the 7% was in like Egypt and Morocco?
      • A: Yes, exactly, super challenging market. That's correct.
    • “[For local-to-local transactions] Adyen was much closer to DLO in terms of price. They, I recall, were in the 4.5% range in Thailand versus around 4% for DLO.”
    • “Yes. Number one, frankly, was regulatory compliance. Two was around flexibility in payment option. And third was around price. And as I evaluated the options that existed here, found that as my team was going through a more detailed analysis of those three. And assessed them through a quant research overall as well as qualitative research from some of their customers as we were talking to them over time, found that dLocal had a chance to beat the rest primarily on the first two. Price, they were mostly competitive, if not a little bit more expensive, but they beat the others on the first two.”
  • Managing Director of a consultancy for payments: “So [Dlocal will] charge up to 5%. That's up to 500 basis points and I think the lowest I've seen is 100 basis points. But say, the lowest is 150 basis points for the larger deals, which is aggressive, compared to EBANX, that's aggressive. EBANX, I have not seen anything lower than 2%, 200 basis points. And so and the take rate, it's a significant part of their margin. Because on the pay-in, they have to buy processing from somebody else because they're not the processor themselves. So I would say 3/4 of the cost, as 60% to 75% of those costs are actually costs to subcontractors. So yes, there is margin, but not as high as on the FX. And of course, if there are predictable volumes, they can also hedge some of the risks away.”
  • Head of Billing & Payments at Customer 4:
    • “But going back to [Dlocal] pricing, again, we have a good commercial agreement with them.”
    • ‘If you're a small merchant, again, we are open to, let's say, this 25 different currencies. The average markup is, let's say, 3%, okay, across the board [net take rate]. For us, it doesn't work like that. Because the size of the traffic that we are maintaining, that is the ability to, okay, I want for that specific currency much better rate or don't add this markup on that, we have more options, right? Because in each and every territory, we're working with at least two partners. So you have a way to find the most attractive offering with the right partner”
    • [Dlocal’s FX markup] “usually, it's between 2% to 5%.”
  • Former Global Lead, Local Payment Methods at Customer 5:
    • “You've got how much the provider takes their cut and then you've got the method cut. And also that's less than 10% maximum, depending on the method but the percentage of the provider is very small. So you're not saving that much really, maybe 1.5% or 2%. If you consider the full cost of a transaction, the bid that is owned by the PSP is not huge. It's maybe in the range of, depending on the method, 1% to 3% or whatever, maximum, let's say, average of 2% [net take rate]. And then the method itself takes a cut of maybe 5% to 6% or whatever.”
    • “We weren't optimizing for how much we pay the provider. Obviously, we were mindful of it, and we took care of [company] expenses around this, but we really, really cared about advertisers. We really cared about how many advertisers we've got on our platform because we knew that once you were in and you were advertising and you saw the value of it, you would continue to advertise on Facebook. So the cost of acquisition is so high that if I'm paying dLocal 2%, 3% or 4%, I mean, yes, 4% is a lot, but it's nothing compared to how much I'm getting from an advertiser.”
  • Head ePayment Analyst at CoralPay Technologies: “The average daily transaction with PayU is about 8-10 Million Nigerian Naira (~$20-26k), and on every transaction, they take a 1.0-1.5% fee, but there’s a cap and the highest fee amount they can take off a single domestic transaction is 1,500 naira (~$4). International transactions range from 3-5% depending on the category.”
    • PayU is a competitor of Dlocal
• Dlocal wins business because of its single API servicing numerous countries such that for small countries it can charge a premium compared to point solutions who only cover that one country. 
  • Former General Manager of the Hospitality Solutions at Customer 3:
  • Q: So it's interesting. I thought I was seeing some numbers I thought suggested 2C2P [a competitor], for example, in Thailand was charging like 1.5%.
    • A: It was around 2% from what I recall.
  • Q: And what was dLocal charging in that same region?
    • A: dLocal's were, I believe, in the range of 3.5% or 4% in Thailand. It was double if I recall correctly.
  • Q: And even still, you guys wouldn't use 2C2P there, why?
    • A: It's the issue with the integration, Thailand wasn't a big enough market for us to justify having a point vendor be there…

Concerns about Client Funds Segregation:

• MW’s analysis of DLO’s Merchants Funds is flawed for many reasons. The analysis shown doesn’t actual imply merchant client funds are missing, the pre-IPO dividend claim doesn’t make sense, and MW does not seem to understand how merchant client funds work.
• MW Report: “Our initial concerns about using client funds arose from the fact that DLO’s 2020 cash flow – the year it paid a $15 million special dividend – does not reconcile. We showed in our report that there was a $3.3 million deficit in DLO’s ability to fund the dividend at the consolidation level.”
  • DLO paid its $15M special dividend to pre-IPO shareholder during Q3’20.
    • Dividend likely occurred before the 8/10/2020 private placement of secondary shares.
  • The dividend was paid >4.5 months before the “so called $3.3M deficit.”
    • The dividend paid has nothing to do with change in trade receivables/payables.  
• DLO Financial Statements:
  • DLO Cash flow statement from Q2’21 – shows no dividend paid YTD through 6/30/20
  • DLO Cash flow statement from Q3’21 – shows $15M dividend paid YTD through 9/30/20

• “Merchant Client Funds” settle relatively quickly, usually within a week, such that it seems quite unlikely that these funds could be used to pay a dividend >4.5 months before without any merchant noticing. DLO’s own financials imply the avg number of days DLO holds merchant client funds is about a week (5-8 days).

• • Former Dlocal Manager: DLO probably couldn’t have even had the time to mix those merchant funds with their own… the merchants always wanted the cash back to USD/EUR as soon as possible… merchants were the first ones to track the cash… couldn’t see how DLO could mix funds…merchant funds are usually settled in a few days to a week… for example, merchants wanted the cash back every-day from Argentina.
• DLO is also known for far faster settlement compared to Ebanx who historically only settled once a month.
  • Former VP EMEA at Ebanx: Dlocal settlement was probably around three days…and we couldn't meet their payout speed… “Generally speaking, yes [dLocal settlement time was faster]. I know that the larger clients at EBANX have, I think, weekly settlement. But generally speaking, EBANX settles once a month.”
• MW’s claim that DLO stole $3.3M in merchant funds also makes no sense when you consider that DLO had $43.7M in its own cash balance and $8.3M in financial assets ($52M total) at the end 2020.

• MW’s analysis of trade’s receivables vs. trade payables doesn’t imply a cash deficit in merchant funds but MW is essentially just calculating the year-over-year change in DLO’s net trade operating working capital (i.e. net trade working capital excluding clients funds).
  • MW Report: “DLO appears to have ended 2020 with $3,347 more cash than the adjusted cash flow statement implies.”
    • Yes, since you basically calculated the change in working capital in 2020. DLO’s net trade working capital had a cash flow benefit but so did many other PSPs such as Adyen, Nuvei, etc. This doesn’t show that funds are missing.
    • In 2021, DLO had a net trade working capital outflow. Again, this is normal and does not imply that merchant funds were “stealing cash” from DLO.
• Let’s look at some simple accounting:
  • Cash Balance of Merchants Client Funds is listed as an asset in DLO’s financial statements.
  • Since the cash is in a segregated account, DLO should have zero balance sheet equity related to that asset.
  • Thus, there is a net trade payable, a liability, that essentially called “net payable due to merchants” that is equal to the cash balance of merchant clients funds.
    • To get DLO’s operating working capital you need to back out the “net payable due to merchants”.
• Below is a simple breakdown of DLO’s 2019/2020 balance sheet between Merchant and DLO trade asset and liabilities.

  • The implied net trade payable (in red text below) should be same as the merchants client cash balance such that merchant net equity is zero

  • DLO’s operating net trade working capital went from $2.3M in 2019 to -$1.1M in 2020, so a $3.4M cash inflow (higher net working capital = working capital investment).
    • MW includes net exchange difference and part of FX changes, such that change was $3.3M.

• Below is an easier way to look at the change operating working capital from 2019-2021.

• MW’s analysis wouldn’t prove even if merchant funds were taken.
  • Even if $5M of merchant funds were stolen this analysis wouldn’t prove it.
    • If I assume that $5M of Merchant Funds were stolen (see below), MW’s analysis of DLO’s net trade working capital would still show an increase of $3.4M ($2.3M to -$1.1M).

• Analysis is the same for DLO’s Malta entity where MW claims the Malta Operating ended with $4,096 more cash than it should have in 2020.
  • The change in net working capital and is shown in the footnotes.

2020 Malta Operating financials:

2019 Malta Operating financials:

• The 2019 Malta financial statements show the breakdown in merchant client funds on asset side (in cash) and liabilities side (inside trade and other payables).
• DLO management could produce disclosure (annual for 2019/2020/2021) that splits out “trade payables” into “net payable due to merchants” and “operating trade payable.” This would definitively prove MW wrong.

Acquisition of PrimeiroPay:

• Nothing is wrong with the PrimeiroPay acquisition by DLO.
  • The timing or accounting of the acquisition is not strange, the shareholders of PrimeiroPay do not seem to be connected to DLO’s management team and seems like solid acquisition of an asset that generated real revenue and gross profit.
• PrimeiroPay was acquired by DLO for $38.67M (+$0.665M earn-out).
  • On March 11, 2021, the contract was signed and it seems like the money sometime between March 19^th (given PrimeiroPay S.à r.l. filing that day) and the end of March.
  • DLO essentially bought PrimeiroPay for its customer base, which is why it was structured an asset acquisition even though it brought incremental revenue and gross profit.
  • Transfer agreement for merchants started on April 1^st.

424B1 Document, 6/4/21

• PrimeiroPay is a real business that generated real TPV/revenue/gross profit from key enterprise merchants.
  • From PrimeiroPay’s Luxembourg subsidiary (PrimeiroPay S.à r.l.) filings you can see that in 2018 it generated €34.6M in TPV and €2.0M in gross profit, which happens to be a 5.78% net take rate.
    • And that PrimeiroPay’s TPV grew 30.7% in 2018.
  • At average 2018 EUR/USD rate of 1.1811x, that implies $2.3M in gross profit in 2018.
    • Assuming a 20-30% 3-year CAGR growth for PrimeiroPay, 2021 gross profit could have been $4M-$5M. So DLO essentially bought out a quality competitor for 8x to 10x current year gross profit.
    • Since PrimeiroPay was only in Brazil/Mexico, this also excluding any revenue synergies from DLO cross-selling other geographies to those new global merchants.
• PrimeiroPay got great reviews from its customers:          
  • Former General Manager of Americas at Customer 6: “So with dLocal, we were very inclined to close the deal with them. They performed really well. However, PrimeiroPay had an advantage, so they sat on the Adyen  network, right? They were a cross-border alternative to access Adyen if you wanted to do Adyen, right? So we were processing Adyen in a lot of European markets, so we say, hey, we could use this gateway over here and just make the calls and have business up with PrimeiroPay pretty fast. So we decided to go down with PrimeiroPay because of that technical advantage. And then when we finalize a trial with PrimeiroPay, we were like, hey, okay, this is good. We like this. We're going to stay with PrimeiroPay. So PrimeiroPay definitely delivered on their metrics and their promises. And we were quite happy. They were definitely approving at a higher rate. They had good account management. They had the levers to build some tools to deliver to clients like Wish, but they definitely met those milestones. Yes, I'm not surprised they got acquired by dLocal. They were delivering quite well.”

PrimeiroPay S.à r.l. 2018 annual report

• PrimeiroPay’s shareholders don’t seem to have any overlap with DLO’s management team,
  • Paired Minds Ltd (904 shares), Peak Data Consulting Ltd (3,070 shares), and TimSuLtd (126 shares)
  • Paired Minds Ltd is likely the holding company of Tim Werner, the co-founder and CEO of PrimeiroPay.
    • PrimeiroPay Blog: “The people in PrimeiroPay are empowered to shape the company’s future with what they can do best. Tim Werner, our founder and CEO, always says that work must be fun and, honestly, the company lives up to this promise.”—
    • According to LinkedIn, Tim Werner continues to work at DLO as General Manager of APAC.

PrimeiroPay S.à r.l. notes of shareholder meeting right after DLO’s acquisition offer (03/19/2021)

PrimeiroPay Technology GmbH filings

Loan Agreement with Insiders:

• There is nothing out of the ordinary about the loan agreements DLO had with Singer/Kanovich. There is nothing inherently wrong with them and occurs at pretty much every venture backed pre-IPO company.
• The facts were never changed. The disclosure MW points to on 6/4/2021 is the same disclosure that DLO had since its DRS/A on 4/13/2021, when its 2020 financials were first filed.
  • Below you can actually see that Singer/Kanovich where at first issued stock options on January 31, 2020 (30,918 options for $1,019 per old DLO share) but later that year when it became clear that DLO was likely going to IPO they took a loan from the Company to exercise those options for tax purposes.
• As stated from the beginning, these loan under IFRS are considered a stock option until the loan is repaid.

DLO DRS/A (04/13/2021)

• The two wire transfers MW showed on December 17, 2020 were not Singer/Kanovich repaying the loans (which would exercise the options) but are inter-company transfer from DLO’s Citi account to DLO’s Safra National Bank account. This wire transfer was likely made to make the loans official for tax purposes but no money actually changed hands between DLO and Singer/Kanovich.
  • The ordering party (account sending the money) for both wire transfers was the same “Dlocal Group LTD” account at Citibank and were sent just two minutes apart from one another. These wire transfers was not money personally from Singer or Kanovich.

• • The beneficiary (account getting the money) for both wire transfers was a Dlocal account at Bank J. Safra Sarasin. This transfer happened frequently enough it was saved as “Dlocal Group Safra” as a name template for DLO’s Citi account.

Subsidiary Accounting Adds Up:

• Dlocal’s three biggest subsidiaries (UK LLP, UK Corp and Malta Operating) can explain ~90% of Dlocal’s IFRS revenue/COGs/Opex and EBIT for both 2019 and 2020. As well as 95.5% of Adj. CFFO in 2020.
  • Although the accounting is slightly different and missing some big subsidiaries (US and Uruguay) these three subsidiaries explain DLO’s financials.

Catalyst

Removal of Muddy Waters short report overhang

Continued growth and FCF generation