At 09:47 on the morning of October 26, 2022, a research note from Gavekal Research crossed Brazilian trading desks with a headline that read like a shrug: the local market may sell off at the open, and you should buy it. The analyst, Rohan Daswani, was not arguing that Brazil had suddenly become cheap in any fundamental sense. He was making a plumbing argument. The selling, in his framing, would come from positions being unwound rather than from fundamentals being repriced, and position-driven selling reverses faster than fundamental selling does.
Four days later, on October 30, Brazilians voted. Luiz Inácio Lula da Silva beat Jair Bolsonaro 50.9% to 49.1%, the tightest runoff since the country's return to democracy. That date gap is the first thing that should bother you. If the note was genuinely written on October 26, it was a pre-positioning call ahead of a vote nobody had priced with confidence. If the date is a recording artifact, the call was published after the fact and dressed as foresight. Either way the analytical frame survives. The confidence attached to the phrase "the market has already priced it" does not.
Here is what almost nobody looked at that week, and what I have been staring at since 2017: the BRL/USDT order book.
In the seventy-two hours before the Brazilian equity market opened on the Monday after the runoff, the spread between the exchange rate implied by local stablecoin order books and the interbank spot rate widened in a way that had nothing to do with Bitcoin's price. Traders were not buying crypto. They were buying dollars through a rail that never closes. That gap is the fastest sovereign-credit print available in the world's most crypto-native emerging market, and it moved before the DI futures curve did. The chart whispers before the market screams.
What follows is not a defence of the Gavekal call or a refutation of it. It is a reconstruction of what that trade actually was, why the crypto market is a better instrument for measuring it than the equity market, and where the argument breaks.
The Setup Nobody Bothers to Explain
Gavekal is a macro research house with roots in the Gave family's Asia-focused work, and its reputation rests on frameworks about willingness to pay rather than on momentum. Daswani's Brazil work fits that lineage. The note's central claim was mechanical: Brazilian assets had rallied hard into the first round on the assumption that Bolsonaro would outperform the polls, that assumption had been validated on October 2, and by late October a portion of that positioning had become crowded. When a crowded position meets an uncertain binary event, the desk-level response is to reduce, not to re-underwrite. That reduction is what produces a sell-off at the open, and that reduction is what reverses within days once the event resolves.
Strip out the jargon and it is a statement about who owns the marginal Brazilian asset. Foreign flows set the price at the margin in Brazil in a way they simply do not in a market like India or China, where domestic institutional pools are deep enough to absorb external selling. Brazil runs a free float, an inflation-targeting central bank, and a domestic pension complex that is large but not large enough. So when global funds trim Brazil, the currency and the curve move together, and they move fast.
Now the number that makes the whole thing work. In October 2022, the Selic rate stood at 13.75%, held there since August after the central bank had hiked from a 2% floor starting in March 2021. Headline IPCA inflation had peaked above 12% in the spring and was decelerating, which means the ex-ante real policy rate was sitting somewhere in the high single digits. That is not a normal carry environment. It is a market demanding compensation for something.
The question that determines whether "buy the dip" is a trade or a trap is what that compensation is for. Daswani's note reads it as a valuation cushion, a discount rate so high that equities and the currency are protected on the downside. That reading is coherent. It is also only one of two readings, and the second one is the one that keeps me up at night.
Before going further, a word on why a crypto strategist is writing about this at all. Brazil is not a side story in digital assets. Chainalysis put the country's on-chain received value at roughly $150 billion across the twelve months to mid-2022, and Brazil has sat consistently inside the global top ten of its adoption index. The BRL is the largest fiat on-ramp in Latin America. Mercado Bitcoin, Foxbit, NovaDAX and the local books of global venues quote BRL directly against BTC and USDT. And in December 2022, the country signed Lei 14.478, which for the first time gave virtual assets a statutory home and designated the central bank as the regulator.
That regulatory move deserves a sentence of context, because it is routinely misread as enthusiasm. Licensing regimes are almost never about embracing innovation. Hong Kong's virtual asset regime was about taking Singapore's Asian hub role, and Brazil's Lei 14.478 was about formalizing a market that had already become the continent's largest before the state had a vocabulary for it. The motive is jurisdictional positioning, not ideology. That distinction matters when you try to read policy risk into Brazilian crypto.
What 'Positions Unwound' Actually Means
There is a difference between a market that is being sold because the story got worse and a market that is being sold because someone needs the cash. The first is informational. The second is mechanical. The first leaves a mark that lasts. The second leaves a hole that fills.
Daswani put the October selling in the second bucket, and on this specific point he was almost certainly right. The evidence is structural rather than statistical. Brazilian equities had run into the first round with a clear directional bet on Bolsonaro beating the polling spread, that bet paid, and the natural response of any fund that had expressed it is to take some off the table before a coin-flip runoff. Nothing about Brazil's fiscal trajectory changed between October 2 and October 26. What changed was the size of the position relative to the size of the uncertainty.
We trade the panic, not the price. A panic built on position reduction has a half-life measured in days. A panic built on a revised growth path has a half-life measured in quarters. Conflating the two is the single most expensive error in emerging-market macro.
I learned the mechanical version of this lesson the hard way in the summer of 2020, farming liquidity on Uniswap V2 with a group of traders in a Discord channel that never slept. I was fast, I was loud, and I was sure. Then I published a guide and, in my own test wallet, left a slippage tolerance set from a previous, thinner pool. The position executed. It just executed into a worse price than the thesis assumed, and the difference ate the entire expected yield plus change. Nothing about my read on the protocol was wrong. My read on the plumbing was. That is precisely the distinction between a position unwind and a fundamental repricing, and it is why I now build a risk footer into every signal I ship.
So far, so clean. The mechanical reading holds. The problem is the sentence that comes next in the argument, because the note cannot resist claiming more than the plumbing supports.
Two Statements That Cannot Both Be True
The note argues that the market had already priced a Bolsonaro victory and, more importantly, had already priced the fiscal discipline that a Bolsonaro government would imply. If that is true, the risk premium embedded in Brazilian assets should have been compressed by late October. Compressed risk premia show up as lower real rates, tighter credit spreads, and a stronger currency.
But the note also describes Brazilian interest rates as unusually high, and treats that as the reason the trade is attractive. Those two claims are load-bearing in opposite directions. If the market had genuinely priced fiscal discipline, the real rate would not be sitting in the high single digits with inflation decelerating. If the real rate really is that high, then the market was never convinced about fiscal discipline in the first place, and the thing you are buying on the dip is not a cheap growth asset.
See the pattern before it prints. This contradiction is not a footnote. It is the whole trade. What you are actually long, when you buy Brazilian risk in that configuration, is a sovereign credit that has been paying you a large coupon to hold it. That is a carry trade. It has been dressed as a structural long.
And a carry trade has a specific failure mode, which is that it pays you a little for a long time and then takes it all back in one afternoon. Anyone who held Turkish lira carry in 2018 or Egyptian pound carry in 2023 knows the shape of the payoff diagram. It looks like a bond until it looks like an equity option that you sold.
There is a way to test which interpretation is right, and it does not require a Bloomberg terminal. It requires watching what the market does with the one variable that carries no ambiguity: the fiscal rule itself.
The Anchor Is the Trade
The Brazilian fiscal story is almost never about the deficit ratio in the year you are looking at. It is about the credibility of the rule that governs the next five years. Between 2016 and 2022, that rule was the teto de gastos, a constitutional ceiling that limited federal spending growth to inflation. The teto was never popular. It was credible, and credibility was the entire point. It functioned as a nominal anchor the way an inflation target functions as a nominal anchor, and Brazilian assets traded off its survival probability rather than off any single year's primary balance.
When Lula won, the market's question was not how much he would spend. It was whether the teto would survive the transition, and if not, what would replace it and whether the replacement would have a hard constraint. That is why the PEC da Transição mattered more than any budget line. The constitutional amendment, negotiated through November and December 2022, carved roughly R$145 billion outside the ceiling for 2023 plus additional room tied to the expanded Bolsa Família program. It was not an enormous number in absolute terms. It was a signal that the ceiling was negotiable, and a negotiable ceiling is not an anchor.
Liquidity is the only truth that bleeds. The market does not price your intentions. It prices the cost of exiting your paper. And the moment the exit cost becomes a function of a rule that can be renegotiated by a coalition, the paper is no longer a growth asset. It is a claim on a political equilibrium.
This is where I part company with the dip-buying frame, and I want to be precise about why. Daswani's argument that a conservative Congress would act as a brake on fiscal expansion is directionally reasonable. The 2022 legislature was the most right-leaning in years, and it did constrain the worst-case scenarios. But the argument treats that brake as free. It is not free. It has a price tag, and the price tag is denominated in budget.
The Brake Has a Price Tag
Under Brazil's coalitional presidentialism, an executive that does not command a legislative majority purchases votes with budget execution. The instrument is the parliamentary amendment, and in 2022 the most consequential version was the RP9 amendment, the so-called secret budget, which routed billions of reais through a reporting channel that did not identify the requesting legislator. Estimates put the 2022 volume in the range of R$16 billion to R$19 billion, and the mechanism was later suspended by the Supreme Federal Court in December 2022.
This is the blind spot in the "Congress will stop him" thesis, and it is not a small one. If the legislature must be compensated for providing fiscal restraint, then the restraint is purchased with fiscal expansion. The net constraint is smaller than the gross one. You can have the brake, or you can have the spending, and the political system will happily sell you the appearance of the brake while booking the spending.
There is a second layer that the note does not touch and that I think is more important than either the deficit or the brake. Brazil's monetary transmission is unusually fast. Floating-rate debt is a large share of the domestic federal public debt, the average maturity sits under four years, and LFTs, which are explicitly tied to the Selic, make up roughly a third of the stock. When fiscal risk rises, it does not show up as a slow deterioration in a five-year projection. It shows up within weeks in the cost of rolling the debt.
That is a double-edged structure. It makes the central bank unusually effective at fighting inflation, because rate hikes transmit into household and corporate cash flows almost immediately. It also makes the sovereign unusually exposed, because the same transmission channel carries a fiscal risk premium straight into financing costs. Brazil has very little tolerance for a policy mistake, and the market knows it, which is why the market prices fast.
Which brings us back to the order book.
The 24/7 Basis
Every Brazilian crypto exchange quotes BRL directly against USDT. That quote implies an exchange rate. The interbank market quotes USD/BRL. The two should converge, and most of the time they do, within a spread that reflects fees, liquidity and settlement friction.
When they do not converge, the gap is information. A positive basis, where USDT costs more reais than the dollar does, means local holders are paying up for dollar exposure through a rail that does not require a bank, a broker, a market maker or a settlement window. During stress that basis widens, and it widens at two in the morning, when the interbank market is closed, the equity market is closed, and the central bank's swap auctions are not running.
The Banco Central do Brasil has a traditional circuit breaker for currency stress: the swap cambial auction, which provides hedges without drawing on reserves. There is no equivalent for the crypto basis. There is no market maker of last resort on a local order book. There is no circuit breaker. There is no closing bell. Which means the crypto basis is the only continuously trading, publicly observable quote on how badly Brazilians want to leave the real.
Speed is the new currency of trust. I have built and discarded more monitoring scripts than I can count, but the one I have kept running since 2021 does one thing: it compares the BRL/USDT implied rate across three local venues and two peer-to-peer books against the last interbank print, and it flags divergence beyond a rolling band. It is not a trading signal. It is a thermometer. In October 2022 it read hot before the equity open did.
I want to be careful here about what this does and does not prove. It does not prove that crypto traders are smarter than equity traders. It proves that the crypto rail has structural properties, no settlement delay, no market hours, no central bank backstop, that make it gap first when the underlying question is about confidence in a currency. The equity market has to wait for the bell. The DI curve has to wait for a session. The order book does not wait for anything.
That is the part of the Brazilian election trade that the macro research never captured, and it is the part that a crypto-native reader already knows in their bones.
What the Dip Is Denominated In
Ask a simple question that almost nobody asked in October 2022: buy the dip in what?
If the answer is Brazilian equities, then the trade is a bet that earnings hold up and the discount rate falls. Both of those are plausible, and both are slow. If the answer is the real, then the trade is a bet that the currency's carry exceeds its fiscal risk premium. That is a different bet with a different payoff shape. If the answer is a dollar-denominated investor buying Brazilian risk, then the trade is a bet on the currency and the asset simultaneously, which means you are long a carry trade with double leverage.
And if the answer, for a Brazilian saver, is stablecoins or Bitcoin, then the trade is not an investment thesis at all. It is a hedge against the exact variable that the dip-buying thesis is long.
This is the contrarian angle that the sell-side framing cannot see, because it is looking at the wrong investor. For a global fund, a Brazilian sell-off is an opportunity to add risk at a better price. For a Brazilian household, a Brazilian sell-off is a reminder that the local unit of account is a political variable. The two investors face the same chart and want opposite things from it.
That divergence has been visible on-chain for years. Brazilian stablecoin volume is not primarily speculative. It is a savings and payments rail, and it grows when confidence in the real falls. The same dynamic that makes a dip attractive to a foreign fund makes an exit attractive to a local holder. Both are rational. Both clear through the same order book. And the order book, unlike the research note, has to reconcile them in real time.
Pixels hold value when code forgets. The irony is that the most reliable fiscal-credit indicator in Brazil is not published by the central bank and is not covered by any sell-side desk. It is a number that thousands of people look at every day without realizing what it means, because they are looking at it as a price rather than as a signal.
There is a related structural point that gets lost in the Layer 2 discourse. The infrastructure Brazilians actually touch is a custodial order book with a fiat on-ramp, not a rollup. Decentralized sequencing has been a slide deck for two years, and it does nothing for a user trying to move out of BRL on a Sunday night. The rails that matter in an emerging market are the ones that settle when the banks do not, and those rails are almost always centralized, custodial and boring. That is not a criticism of the technology. It is an observation about which layer of the stack carries the actual risk.
Duration, Not Value
Let me put the whole argument in one frame, because the frame is the contribution.
The Brazilian dip-buying trade in late 2022 was a duration trade wearing a value trade's clothes. You were being paid a high coupon to hold a claim whose price depended on the credibility of a fiscal rule that was in the process of being replaced. That is duration risk. Duration risk pays well until the rule changes, and then it pays nothing at all.
Chaos is just data waiting to be decoded. The decoding key is not the deficit number. It is the transition period, the window between the old anchor being abolished and the new anchor being believed. That window is when the risk premium is widest and the information is thinnest, and it is exactly the window that a speed-first strategist should be watching rather than a value-first one.
I have made this mistake in its purest form. In the back half of 2022, sitting in a bear market with positions bleeding and no clean read on the structural damage inside failing lenders, I spent more nights in poker games with other traders than in front of a curve. I published calls that the bottom was near because the room felt like the bottom was near. The room was wrong. The data was available and I was not reading it. The lesson was not that sentiment is useless. It was that sentiment is a lagging indicator of positioning, and positioning is a lagging indicator of the thing that actually kills you, which is a rule change you did not model.
So here is the discipline I apply now. Any time a market is offering me a real yield in the high single digits, I ask three questions in order. What is the rule that governs the repayment? Is that rule being replaced right now? And what is the fastest observable price of that rule changing? For Brazil in 2022 the answers were: the teto, yes, and the BRL/USDT basis. That is a complete chain, and it does not require a view on Lula or Bolsonaro at all.
The Signal That Keeps Printing
The Brazilian election is over. The framework it revealed is not.
The teto was eventually replaced by a new fiscal framework in 2023 that ties spending growth to revenue growth within a band, and the real went on to be one of the better-performing major currencies of that year as the carry reasserted itself. The dip, in other words, was buyable. Daswani was directionally right. I want to be honest about that, because a contrarian argument that refuses to concede a correct call is just contrarianism.
But being right on direction and right on mechanism are different things, and only one of them is repeatable. If you bought the dip because positions were unwound, you made money and you learned nothing. If you bought it because you believed the fiscal anchor was intact, you made money and you learned something false, which is worse.
The thing to watch going forward is not the headline deficit. It is the hard constraint. A fiscal framework with a binding, non-negotiable ceiling is an anchor. A framework with an escape clause that a coalition can invoke is a schedule of future negotiations. Brazil's new rule sits somewhere in between, and the market's job for the next several years is to price which side of the line it falls on.
And watch the basis. Not the equity open, not the DI curve, not the central bank's communiqué. Watch the gap between what USDT costs in reais on a local order book and what the dollar costs in the interbank market. When that gap widens on a Sunday night, someone with a balance sheet is telling you something about the real that no research note will say out loud until Monday.
The question worth sitting with is not whether Brazil is cheap. It is whether the compensation you are being paid for holding Brazilian risk is a discount rate or a warning. For most of the last decade it has been both at once, and the only way to tell the difference is to find the fastest price of the rule and read it before the bell.
