Can You Make Money Day Trading? Brazil Checked

Of the fifteen hundred and fifty-one Brazilians who day traded on more than three hundred days, eight earned more per day than a bank teller's starting pay. Not eight percent. Eight people. So can you make money day trading? Brazil checked. Its regulator's records covered everyone who started day trading in one market over three years. Of those who kept going past three hundred days of trading, ninety-seven percent lost money.

You've heard the other version, and it sounds reasonable. Trading is a skill like any other, so learn the strategy, master your emotions, put in the screen time, and you can do it for a living, in stocks or in crypto. A popular finance website told its readers in twenty nineteen that anyone with ambition and patience can trade, and do it for a living. The same site said reaching a consistent income would likely take a year or more, full time. Everyone in that ninety-seven percent had day traded on more days than the market is open in a year. And the longer people kept at it, the fewer of them were in profit.

If you've ever thought about trading your way out of a job, or you know someone who's trying, this crowd is worth following. It isn't a handful of volunteers or a poll. It's everyone who started, in one market, over three years. So we're going to follow nineteen thousand six hundred and forty-six people from their very first day trade. How many of them lasted? How many made money? Did the ones who stayed get any better? And is anyone actually good at this, or is every winner just lucky? There's also one kind of trader none of this data reaches, and we'll name it before the end.

First, the believer's case, because it rests on real numbers. In any single year, plenty of day traders do make money. On Taiwan's stock market, about one in five of the more active day traders finished a typical year in profit after fees. Older studies summarised by the Brazil researchers found about three in ten in Finland, and just over a third in a group of three hundred and twenty-four American traders. So if you know a day trader, there's a fair chance you know a winner, and that winner isn't lying to you. Why would you doubt someone who shows you a profitable year?

The second half of the case is about who stays. In Taiwan, more than three in four day traders quit within two years, and the ones who lose tend to quit faster than the ones who win. Read that the way a believer would, and it sounds like a filter. The market shakes out the people who can't do it, and the ones still standing after a year or two are the ones who can. Practice helps as well, at least for some: among Taiwanese traders who were already profitable, results improved with experience. If that filter works, the crowd that's left should be full of winners.

Brazil is where that filter got tested. The country's securities regulator handed three economists the daily trading records of every individual and institution in one market, each person anonymised but tracked over time. The market was mini index futures, which the researchers call the preferred asset of Brazil's day traders. A future here is a contract that rises and falls with the stock index, so you can bet on it going up or down inside a single day. From those records, the researchers pulled out everyone who day traded it for the first time between twenty thirteen and twenty fifteen. That came to nineteen thousand six hundred and forty-six people, followed from their first trade.

So how many of them lasted? Just over one in twenty day traded for a single day and never did it in that market again. Half of the whole crowd stopped somewhere between their second day and their fiftieth. By the time you reach people who day traded on more than three hundred separate days, there are fifteen hundred and fifty-one left, about one in thirteen of everyone who started. Those are days on which they actually traded, not calendar days, so this is more than a year's worth of market sessions. They're the committed ones, exactly the people the believer's filter should have kept.

Now look at who in that crowd was in profit after the exchange and broker fees. Among the people who traded for only one day, almost three in ten came out ahead. Among those who lasted up to fifty days, it was about one in six. After that the share keeps falling, to about one in eleven, then one in fifteen, then one in nineteen. And among the fifteen hundred and fifty-one who went past three hundred days, it was three in a hundred. The researchers point out that roulette shows the same pattern, where the share of winners shrinks the more rounds you play. One plain reading is that more days give luck less room to hide a small, steady loss. Staying in didn't keep the winners.

How many of those fifteen hundred and fifty-one actually made money? After fees, forty-seven of them made any profit at all, which leaves fifteen hundred and four who lost, ninety-seven percent. Seventeen earned more than Brazil's minimum wage, which the researchers work out at sixteen dollars a day. And eight earned more than a bank teller's starting pay, about fifty-four dollars a day. If you've seen ninety-one percent quoted for this study, that's the same people before broker fees, in the same table. Eight people out of fifteen hundred and fifty-one.

The one in the middle of that group lost about twenty-three dollars for every day they traded, and the average loss was nearly forty-nine. And how did the very best of all fifteen hundred and fifty-one do? About three hundred and ten dollars a day, on average, and the researchers say that came with great risk. That one person's results swung by thousands of dollars from one day to the next. None of these figures include income tax, courses or trading platforms, so the authors say performance is likely over-estimated. That's the ceiling.

The believer has one more answer here: these people were still learning, and the profits come later. The researchers tested exactly that, across more than seven hundred thousand days of individual trading records. Over their first two hundred and fifty days, the people who stayed lost about forty-seven dollars a day on average. After day two hundred and fifty, they lost about fifty-two. In fairness, the very top of the group did edge up a little later on, so a handful may well have improved. Did the ones who stayed get better? On average, no.

Losing and coming back anyway sounds irrational, but in Taiwan it was ordinary. Among traders there with at least fifty days of experience and a losing record, about ninety-five in a hundred were day trading again within a year. Among those with a winning record, it was about ninety-six. The researchers' verdict on how traders react to their own losses is blunt: sensitive to losses, but not very sensitive. A losing record barely changed who came back.

Every trade has somebody on the other side. So who was on the other side of all those losing trades? Brazil's paper doesn't measure where the money went, but it shows who moved in. Firms whose machines close more than one deal a second handled about one in nine of all deals there in twenty twelve. By twenty sixteen it was about four in ten. The more of the market those machines had, the more individuals tended to lose. That's a link in the data, not proof of a transfer. Then, at the end of twenty eighteen, after the years this crowd was followed, brokers zeroed their fees for day traders. The researchers say it was to lure new customers to their courses and platforms. Zero fees were the bait.

On the seventh of March twenty nineteen, Brazil's largest financial newspaper put this study on its front page. Course providers who teach day trading answered on YouTube, in more than a dozen videos, and the biggest of those had over ninety thousand views a few weeks later. Meanwhile the number of retail day traders, which had been climbing fast, broke its upward trend. The researchers are careful to say other things may have contributed. Still, it's the one moment in this story when people saw the odds before they started. The trend broke.

That leaves the question a believer should be asking us by now: is anyone actually good at this? The best answer comes from Taiwan, where economists studied every day trade on the whole Taiwan Stock Exchange from nineteen ninety-two to two thousand six. In an average year, about four hundred and fifty thousand people day traded there. About one in five of the active ones made money after fees in a typical year. But when the researchers asked who did it predictably, year after year, the answer was somewhere between one thousand and four thousand people. Under one percent.

That under one percent is real, and we're not going to wave it away. Taiwan's best day traders, ranked by the previous year's record, kept earning after fees the following year. The authors put it plainly: most of the successful ones are merely lucky, and a small group is skilled. So skill exists. The trouble is that from the inside, a lucky year and a skilled year look exactly the same. The only way to find out which one you had is to keep trading, and Brazil showed what keeping on tends to cost.

And even with that skilled group included, Taiwan's day traders lost money as a group, after fees, in every one of the fifteen years studied. Before costs their losses were small, and trading costs more than tripled them. So who came out ahead year after year? The researchers name the beneficiaries of what they call overconfidence: brokerage firms, through commissions, and the government, through a tax on every sale. The steady winners were the brokers and the taxman.

All of that is stocks and futures, years ago, and crypto people will say their market is different. Here, for once, the trades are public. Hyperliquid is a crypto exchange for perpetual futures, contracts that let you bet on a coin's price with no end date, and with leverage, which means partly with borrowed money. Its orders and trades are recorded on a public blockchain. Four researchers at Stanford and Columbia took that complete record from March twenty twenty-five to the end of June twenty twenty-six, for its ten most traded contracts. They sorted the wallets into three groups: market makers, trading algorithms, and people clicking buttons on the website. Which group do you think paid?

On Hyperliquid, the people trading by hand through the website lost one point one one billion dollars in fifteen months, by the paper's measure. Market makers gained about two hundred and eight million, and the algorithms roughly broke even. Seven hundred and twenty-five million of that loss was fees. Another two hundred and twenty-six million went when leveraged bets lost enough that they were closed out by force. Only about a hundred and fifty million was lost on the trading itself. The authors add that none of this requires trading against better-informed counterparties in the usual sense. Nobody had to outsmart them. About two dollars in every three they lost went on fees.

That billion is measured two hours after each trade, not on what each wallet finally made or lost. And a wallet isn't a person, so this is a group's total, not the share of traders who lost. Nobody has yet followed crypto day traders from their first trade to the day they quit. The crypto data counts dollars, not careers.

The fairest objection to the Brazil and Taiwan numbers is that they're old. Aren't the tools and the teaching better now? Two of the same Brazilian researchers looked again. From the eleventh of March twenty twenty, when the pandemic was declared, to the end of twenty twenty-three, nearly nine hundred and seventy thousand people day traded futures in Brazil. At the peak, it was around a hundred thousand people a day. Between them they lost nine point nine billion reais. On average that's about ten thousand reais each, though some made money, and it's before a single fee or tax.

In one well-known economic model of day trading, the typical beginner is a hairdresser who trades between haircuts. In Brazil's pandemic data, the hairdresser stopped being a model. About six thousand real hairdressers day traded futures, and on average each lost about five and a half thousand reais, before costs. Doctors lost about thirty-five thousand each, on average. Not one occupation in the data came out ahead on average. Is that the same as a crowd followed from the first trade? No, and we won't pretend it is. But it doesn't look like the problem went away.

The objection you'll hear most is that these were amateurs, and that real traders use a strategy, risk management and discipline. Part of that we can answer. The fifteen hundred and fifty-one weren't tourists: they kept coming back for more than a year's worth of sessions, and on average they were no better after day two hundred and fifty. In Taiwan, losers came back almost as often as winners. But we can't see inside anyone's head. Can a trading record show who took it seriously? It can't.

A Brazilian lawyer with a postgraduate specialisation in finance wrote a long rebuttal to an earlier version of this research, and his best point is exactly that one. The data, he wrote, doesn't show how many people did it seriously, by taking courses, reading, training and respecting methods. He's right, and we can't answer it. He also argued that judging traders on their first years is like judging doctors by their residents, and nobody in Brazil's data is a ten-year veteran. His verdict is that it is possible to develop the skills to profit consistently. Yet in the same passage he concedes that success comes to only a small fraction of investors, leaving, in his words, a trail of destruction for the others.

Now the trader we promised at the start, the one none of this data reaches. Some people trade for a living inside firms, or with a firm's money in a funded account. In Brazil's data, an account in a firm's name counts as an institution, not an individual, and none of these studies picks those traders out. Nobody has published their numbers, so we can't tell you how they do. Of every objection to this video, that's the one we answer worst.

Can you make money day trading? Some people do, and Taiwan shows a small group doing it with real skill, under one percent of day traders. But when a regulator's records covered every new day trader in a market, almost nobody who kept at it earned even a bank teller's starting pay. In Brazil, ninety-seven percent of the people who went past three hundred trading days lost money, and on average practice didn't make them better. On Hyperliquid, the people trading by hand lost over a billion dollars in fifteen months, most of it in fees. Skill is real. In this data, it's also rarer than one in a hundred.

If this is right, then the next time someone shows you a winning month, the question isn't whether it's real. It may well be. The question is how many people started alongside them, and where those people are now. Brazil counted that crowd from the first trade. Eight people out of fifteen hundred and fifty-one earned more than a bank teller's starting pay. So is that proof the skill is real, or proof it isn't worth chasing? Tell us which side you're on.