What if you only bought stocks that just issued new shares?

Stocks whose share counts rose over 20 trading days

Listed shares increase when a company issues new shares. The count rises when a company sells new shares through a paid-in capital increase, convertible bonds are converted into shares, or executives and employees exercise stock options. This test looked only at the numbers, without distinguishing the causes. Each Thursday, using closing data, it collected stocks whose share counts had increased by more than 0.1% but less than 50% from 20 trading days earlier. Changes by exact multiples such as 2, 5, or 10, as in stock splits or reverse splits, were excluded because prices change along with them. Ratios corresponding to increments of 0.05 shares per share, from 1.05 times to 1.45 times, were also excluded because they resembled bonus issues. Bonus issues bring in no new money, so existing shareholders' ownership stakes do not shrink. Alteogen's bonus issue of 0.3 shares per share was classified as dilution under the initial definition, so this filter was added. Among the remaining stocks with a market cap of at least 30B won, up to 30 were selected in descending order of percentage increase.

After 68 trading cycles, the account was worth 2.88M won

The account started with 10M won on June 3, 2021. It peaked at 10.06M won on July 29, the second time all holdings were sold. It never rose above 10M won again. Its highest value after that was 9.17M won on September 23, 2021. Near year-end, the account was worth 5.5M won in December 2022, 3.83M won in December 2023, and 2.78M won in December 2024. At the final sale on August 27, 2026, it ended at 2.88M won. An account that bought all KOSPI and KOSDAQ stocks with a market cap of at least 30B won at equal weights on the same dates ended at 7.67M won. This market account also ended below its initial capital. The KOSPI index gives larger stocks higher weights, but this account gave small stocks the same weight as large ones. In 22 of the 68 four-week periods, the account holding stocks with increased share counts outperformed the market account. The test was repeated with 4 accounts whose start weeks were staggered by 1 week. They ended at 1.95M to 3.68M won, while the market accounts ended at 7.59M to 7.67M won.

The gap widened with longer holding periods

Separately from the simulated account, a new list was selected each week. Its returns after 1, 4, and 12 weeks were compared with the market average over the same dates. After 1 week, returns were 0.22%p lower on average, and 40% of weekly lists outperformed the market (t -1.62). After 4 weeks, returns were 1.41%p lower, and the share of weekly lists that outperformed fell to 33% (t -2.79). After 12 weeks, returns were 3.99%p lower, and 26% outperformed (t -3.21). A comparison that randomly selected the same number of stocks from the market in the same weeks was repeated 300 times. Neither the 4-week nor the 12-week comparison produced a result as weak as the observed value even once (random comparison p 0.003). However, 12 weeks was chosen as the main evaluation period after reviewing the results for 1, 4, 8, and 12 weeks. The values for all periods are therefore reported together, rather than only the selected period.

They also underperformed while short selling was banned

Short-selling rules changed twice during the sample period. Through November 3, 2023, short selling was allowed only for KOSPI 200 and KOSDAQ 150 constituents. From November 6, 2023, through March 30, 2025, it was banned for all stocks. It resumed for all stocks on March 31, 2025. Over 12 weeks, returns relative to the market were -3.21%p (t -1.92) during the period when short selling was allowed, -4.78%p (t -2.44) during the ban, and -4.64%p (t -2.03) after it resumed. The stocks underperformed the market in all 3 periods, making it difficult to attribute the result to short selling. The 4-week gap, however, narrowed from -1.86%p during the period when short selling was allowed to -0.73%p (t -0.73) after it resumed. The result also has limits. The test did not distinguish capital increases from convertible bond conversions. It did not deduct the fees and taxes incurred when replacing all 30 stocks every 4 weeks. This is an observation that rising share counts weighed on stock prices over 5 years, not a signal to trade this list in the opposite direction.

Key numbers

Account holding 30 stocks with increased share counts

2.88M won (-71.2%)

Account holding the whole market

7.67M won (-23.3%)

Relative to the market average after 12 weeks

-3.99%p (t -3.21)

Relative to the market average after 4 weeks

-1.41%p (t -2.79)

4-week periods that outperformed the market

22 of 68

4 accounts with different start weeks

1.95M to 3.68M won (market 7.59M to 7.67M won)

FAQ

Were stocks whose share counts rose through bonus issues included?

No. Changes by integer multiples such as 2 or 5, and ratios from 1.05 times to 1.45 times in increments of 0.05, were excluded because they resembled splits or bonus issues. Bonus issues bring in no new money, so existing shareholders' ownership stakes do not shrink.

Why did share counts increase?

The data used for this calculation cannot show that. The Public Data Portal's price data provide listed shares but do not identify whether an increase came from a paid-in capital increase, conversion of convertible bonds, or exercise of stock options. Results could be stronger or weaker when broken down by cause. The cause must be checked in each company's filings.

Would trading costs change the result?

They would worsen the results for the account holding stocks with increased share counts. This account sells all 30 stocks and buys a new list every 4 weeks, so each of the 68 trading cycles incurs fees and securities transaction tax. The gap against the market account would be wider than in these results, which exclude costs.