One documented 95% claim appears in NASAA’s 1999 report as a quotation of Harvey Houtkin, principal of All-Tech. The passage compares day trading with other businesses; it is not accompanied by a trading-outcome measurement. That does not establish the first use or every later source of the number.
Publication check, 15 October 2026. NASAA’s report preserves Houtkin’s 95% quotation, the unattributed 90% passage and a separate small account analysis. This establishes a documented appearance, not the first or only origin of every later 95% claim. Historical studies use different samples and outcomes; none measures a universal 2026 loss rate.
Where it comes from
The trail for this particular 95% claim ends in a quotation preserved in a regulator's document.
In its 1999 Report of the Day Trading Project Group, the North American Securities Administrators Association quotes Harvey Houtkin, principal of All-Tech, from the Securities Regulation and Law Report in February 1999:
"Day Trading is a business like any other. It's not wild speculation. And, like other businesses, 95 percent will fail in the first two years."
Read what he was actually doing. The quoted passage presents day trading as an ordinary business and compares failure rates. That is an interpretation of the words preserved in the report, not a verified account of his intent or a new statistical study.
The 95% is his comparison with other businesses. The quoted passage is not a measurement of trading outcomes, and no trading-data method accompanies it. NASAA did separately analyze a sample of All-Tech customer accounts; that is a different finding and should not be confused with Houtkin's 95% statement.
The report provides a way to trace this particular quotation. A later article’s citation chain must be checked individually before asserting that it derives from the same passage.
Why it spread so effectively
Three properties made it durable.
It is memorable. A round number close to certainty. "95% fail" reads as a law of nature.
It appeared in a regulator's report. Anyone finding the quote inside a NASAA document could reasonably assume it carried official weight. NASAA used the quotation in its discussion, but it was not a measured 95% outcome from NASAA’s own account analysis.
It sounds conservative. Because it is worse than the figures most people expect, quoting it feels like intellectual honesty. Repeating a pessimistic number seems safer than repeating an optimistic one. It is not safer if the number is invented.
What the real research says
The studies below produce different estimates for different populations and outcome definitions; none is the source of Houtkin's 95% statement.
| Study | Market | Period | Finding |
|---|---|---|---|
| Garvey & Murphy (2005) | US direct-access broker | 68 trading days, 2000 | About half of the 1,386 accounts profitable after commissions |
| Jordan & Diltz (2003) | US broker sample | 1998 to 1999 | Roughly two losers to one winner; about 20% more than marginally profitable |
| Barber, Lee, Liu & Odean (2004) | Taiwan | 1995 to 1999 | More than eight in ten lost money in a typical six-month period, under the paper's cost assumptions |
| Chague, De-Losso & Giovannetti (2020) | Brazil, equity-index futures | 2013 to 2015 entrants | 97% of 1,551 traders who persisted beyond 300 days lost money |
| Barber, Lee, Liu & Odean (2014) | Taiwan | 1992 to 2006 | Fewer than 1% predictably and reliably earned positive abnormal returns net of fees |
Notice that these are not competing estimates of the same quantity. They measure different things: any profit, profit net of costs, beating the market, beating it repeatedly, earning a living. The 95% figure sits in the middle of that range by coincidence, not by measurement.
The report’s 90% passage also lacks a named study
The report also contains a separate 90% claim whose quoted passage names no study.
An attributable 1999 appearance is again inside the NASAA report: it records a brokerage advertising an "85 percent success rate" while saying that "some analysts claim there is a 90 percent failure rate."
Unnamed analysts. No study. Reported second-hand from a brokerage marketing page. Quoted in a regulator's document. That is the provenance of the passage documented here, not proof of every use of the number.
That passage does not identify those analysts or provide their method. We have not established that it was the first use of the 90% figure, or that no independently measured 90% result exists for any defined group.
FINRA's Rule 2270 day-trading risk disclosure warns that traders should be prepared to lose the funds they use and says an investment under $50,000 can significantly impair profitability. It does not supply a 90% loss estimate. Do not attribute that number to this FINRA rule.
The figure that is real
One number in wide circulation does trace to serious work: 80%.
Barber, Lee, Liu and Odean, using complete Taiwan Stock Exchange transaction records, found that "in the typical six month period, more than eight out of ten day traders lose money." That is a real finding from a real dataset.
What gets lost is every qualifier: Taiwanese traders, net of transaction costs, over a six-month window, between 1995 and 1999, under a deliberately low commission assumption. Strip those and you have a free-floating "80%" that is applied to American retail traders in 2026 as though the study had measured them.
Why this matters more than pedantry
Three practical reasons.
Fabricated numbers are fragile. If you make a decision because "95% fail," and someone shows you the quote is a firm owner's business analogy, the entire basis of the decision evaporates. A conclusion built on a real finding survives scrutiny.
An unsourced figure is not a conservative estimate. The Brazilian equity-index-futures study found that 1.1% of traders who persisted beyond 300 days earned more than Brazil's minimum wage from this activity. That is a narrow, documented finding, not a universal day-trader success rate.
Bad statistics discredit good warnings. Every time a fabricated number is debunked, it gives someone a reason to dismiss the whole body of evidence. The evidence is strong. It does not need help.
The number, restated properly
Houtkin's analogy dates to 1999. Distinct research findings include: about 20% of Taiwan day traders earning profits net of fees in a typical year in the 2014 study; under 1% of the broader Taiwan day-trader population predictably profitable across years; 97% of the persistent Brazilian index-futures subgroup losing money; and a separate literature on learning and attrition. These studies cover different populations, markets and definitions of success; their results cannot be combined into a universal current day-trader failure rate.
None of those is 95%.
None needed to be invented.
Trace any loss statistic you encounter through the five-step provenance check this page runs — who measured, what population, which market, what window, net of which costs — and the figure either survives with its meaning intact or dissolves into a 1999 sales analogy repeated by people who never read the transcript.
A percentage is useful only with its population, period, outcome definition and cost assumptions intact.
Before you act on this
This article explains how something works. It is general education, not advice about your situation. It does not consider your goals, income, tax position or how much risk you can afford.
Investing involves risk, including losing money. Before you act, speak to a licensed professional. You can check whether someone is licensed at Investor.gov and FINRA BrokerCheck.
Common questions
Is Harvey Houtkin a credible source on this?
NASAA identifies him as All-Tech’s principal and attributes the passage to a February 1999 trade-publication article. He had an industry interest; the quoted analogy is distinct from measured account outcomes.
So is day trading actually risky?
Yes. The cited Taiwan and Brazilian studies show high loss rates under their defined methods and populations. They cannot be turned into one universal rate for every market or time period.
Where did "80% of traders lose money" come from then?
One documented source is Barber and coauthors' 2004 Taiwan study. Its result is for a typical six-month period, net of the paper's assumed trading costs. It is not a current US estimate.
Why does NASAA’s report contain so many of these?
It documented industry claims and criticism during the 1999 day-trading boom. A quotation in the report is not automatically a result measured or endorsed by NASAA.
How to check a statistic like this yourself
The method that produced this page is repeatable, and worth knowing because these figures circulate constantly.
Follow the citation chain to its end, not to the first respectable-looking stop. A blog cites a news article; the news article cites "a study"; the study, when you find it, measured something narrower. Keep going until you reach a document with a methodology section, or until the chain runs out. If it runs out, that is your answer.
Check who is speaking. The 95% figure survives because it appears inside a regulator's report. It is a quotation of an industry principal, not a regulatory finding. Reading one more sentence of context changes what the number means entirely.
Check what was measured. "Fail," "lose money," "are not profitable" and "do not beat the market" are four different tests producing four different numbers from the same population.
Check the horizon and sample. Garvey and Murphy (2005), not Jordan and Diltz, reported roughly half of traders profitable across a 68-trading-day US sample. The long-run Taiwan evidence reports a much smaller persistently profitable group, but the markets, samples, methods and horizons differ, so a “67-fold” comparison is not meaningful.
The numbers in NASAA's report are not interchangeable
The 1999 NASAA report contains three very different kinds of percentage. The 95% is Houtkin's business analogy. The 90% is a brokerage's unattributed reference to unnamed analysts. The 70% is not fabricated: it comes from NASAA's consultant's analysis of 26 accounts at one All-Tech branch, after 30 were sampled and four removed to avoid double-counting individuals. About 70% of those reviewed accounts lost money under that analysis.
The 70% is a genuine, very small, firm-and-branch-specific study result. It cannot be converted into a measured loss rate for all day traders. Nor can the 90% and 95% quotations be treated as if NASAA had measured them. The right question is not whether a number appears in a regulator's document, but what evidence produced that particular number.
Where to go next
If you want to check claims like these yourself, our investing glossary defines the terms these studies actually use — abnormal return, net of fees, persistence — and the distinctions between them are usually where a statistic goes wrong.
Related reading: Where every day trading loss statistic comes from · What the profitable day traders actually look like · Why day trader profitability studies disagree
Sources
Primary sources. NASAA's 1999 report · FINRA Rule 2270 · 2004 Taiwan study · 2014 Taiwan study · Brazilian index-futures study
Education, not advice. Historical results do not predict future returns. Nothing here is a recommendation regarding any security or strategy.
A paper trading simulator can help you practice keeping a trade log. Simulated performance is not a personal success-rate estimate or proof of live-trading skill; fills, costs and behavior can differ.
Additional source detail
NASAA, Report of the Day Trading Project Group (9 August 1999) — contains both the Houtkin quotation and the "90 percent failure rate" reference. Ronald L. Johnson, Day Trading: An Analysis of Public Day Trading at a Retail Day Trading Firm (1999). Barber, Lee, Liu & Odean, "Do Individual Day Traders Make Money? Evidence from Taiwan" (2004) and Journal of Financial Markets 18 (2014). Chague, De-Losso & Giovannetti, "Day Trading for a Living?" Garvey & Murphy, Journal of Applied Finance 15:2 (2005). Jordan & Diltz, Financial Analysts Journal 59:6 (2003). FINRA Rule 2270.
Cite this article: Akbar Shah (2026). 95% of Day Traders Fail? What NASAA’s 1999 Report Actually Says. StockEducation.com.
License. This may be reproduced with attribution to StockEducation.com and a link to this page.
