The Data on Day Trading


Fueled in part by lockdown boredom, unexpected pandemic income, and commission-free brokerages (Robinhood), and then further juiced by an epic bull market and the rise of volatile cryptocurrency trading, day trading has seen a rise in popularity over the last few years. The internet is rife with influencers peddling their particular courses and video series, promising to unlock the secret to windfall profits from only a few hours of daily work.

This post looks at the available data and financial literature on the feasibility of profitable day trading, and the results are awful. Almost all day traders lose money, particularly after accounting for fees and taxes. Astonishingly few individuals are able to consistently profit over time, and those that do are extremely experienced and sophisticated -- not beginners.

While certainly not comprehensive, this post will continue to be updated over time with additional studies.

The Studies

This paper seeks to understand the differentiating factors between successful and unsuccessful day traders in Taiwan. The authors received comprehensive trading data from the Taiwan Stock Exchange from 1992 through 2006. They defined day trading as the purchase and sale of the same security, by the same individual, on the same day, which accounts for approximately 17% of the overall volume on the Taiwan Stock Exchange. On average there were 450,000 individual day traders during each year, of which about half (277,000) traded in amounts larger than $20,000 USD-equiv each day. The key findings were:

  • Any given year, only about 19% of the > $20k day traders made positive abnormal returns (i.e., did better than the market) net of fees. That's 1 in 5.
  • Only 4,000 individuals (less than 1% of the population of day traders) were able to consistently profit, net of fees.
  • The top 500 traders were remarkably consistent and generated outsized profits, earning net +37.9bps per day.
  • The "bottom" 440,000 traders (the overwhelming majority of the 450,000 population) lost about 25-29bps per day.
  • While the overwhelming majority of traders lost money - data showed a consistent trend that highly active day traders outperformed occasional day traders.
  • Trader performance was statistically more consistent and non-independent than random luck would predict. That is, the profitable traders (of which there were exceedingly few) were not just getting lucky.
  • Profitable day traders performed especially well trading few volatile stocks near earnings announcements - suggesting that the source of profits comes from their superior trading judgement and expertise in those stocks, and not from just providing liquidity to the market.
  • Insider trading may be attributable to some of the profits, and cannot be ruled out, though the consistent profitability of the highest performing day traders, even outside of earnings announcement periods, suggests this is not entirely the case.
  • The evidence suggests that profitable day traders "react more quickly to public information signals in their trading strategies" than the unsuccessful traders.

In conclusion, the paper finds strong evidence that day trading is consistently and highly profitable, even net of fees, to those with 1) the best information/strategies, 2) the best systems/access to trade quickly, and 3) the highest conviction/discipline. However, this accounts for fewer than 1% of all traders. In all likelihood, these are career professionals with great connections and highly sophisticated trading tools and systems. The overwhelming majority of day traders lost money, consistently, and worse than random luck would predict.

This study is an analysis of all new day traders on the Brazilian equity futures market from 2012-2017. (Futures are particularly compelling for day traders in the USA as well, as they are highly liquid markets and exempt from particular PDT taxes that apply to day trading most other securities).

The distinction here of only using new day traders is interesting, since it very likely weeds out the highly sophisticated professionals referenced in the Barber study above. Rather, this study is probably a more realistic view of what an 'everyday' person can expect if they quit their day job and start day trading. Highlights and findings from the study:

  • The study covers about 20,000 individual investors who began day trading in 2013, 2014, and 2015, defined as an investor who did not have any day trades in the prior calendar year.
  • Returns clearly show that probability of being profitable goes down proportionally with the volume of trades made. That is, the most profitable cohort of day traders were those who only made a single day trade all year.
  • Of the most active traders (1,500 of the 20,000 population), data shows performance did not improve over time. I.e., those individuals did not learn or improve their skills in subsequent trades.
  • Of the most active traders, only 17 individuals (1.1% of the active group) earned more than Brazilian minimum wage, net of fees.

Overall, the paper concludes that "it is virtually impossible for an individual to day trade for a living, contrary to what brokerage specialists and course providers often claim."

This is a brief study done during the first internet bubble, when online brokerages gained popularity and day trading first because a viable option for regular people. The intent of the study was to determine the overall profitability of day trading in light of its contemporary rising popularity.

The researchers were able to study the trades of 324 day traders from February 1998 to October 1999. Note that this was a time period where the S&P500 gained 30%, and the Nasdaq was up over 100%. Findings and highlights:

  • Of the 324 traders for whom they had data, 36% made a profit over the period of the study, while 64% made a loss.
  • Only 20% of the traders in the study made more than $5,000 over the time period.
  • Total commissions for each round-trip trade were $30. Researchers determined that on average (across all investors) day trades were profitable gross (pre-fees), but were negative net (after fees).
  • Researchers found a statistically significant correlation between the performance of day traders and the price movement of the Nasdaq exchange, indicating that traders had a preference to trade long positions in tech stocks. Again, the study was during a period where the Nasdaq more than doubled - so the result that only ~1/3rd of day traders made any profit at all makes that result seem quite a bit worse.

The study concludes:

The day trading industry, and numerous books and advertisements about day trading, actively promote the idea that day trading is an easy route to wealth and early retirement. This idea has led to widespread interest in day trading and the entry of many novice day traders. The fact that at least 64% of the day traders in this study lost money suggests that it is more difficult to be a profitable day trader than the industry maintains.