Meme Stocks Over the Long Run

Is the meme stock phenomenon new, or is it a persistent feature of stock markets made visible only recently via social media? In his July 2026 paper entitled "A Century of Meme Stocks: The Modern, Ephemeral Premium", Chad Schmerling addresses this question by constructing meme stock selection models based on market data alone and extendable in some form back to 1926. The model specifications measure the extent to which a stock's price is being set by coordinated attention rather than fundamentals. The specifications derive from the holdings of a prior passive incarnation of the Roundhill MEME exchange-traded fund (ETF), which selected stocks based on WallStreetBets mention scores and short interest. He translates these holdings into patterns based on the following data: shape of recent returns; trading volume/abnormality; option surface; option flow; short-sale constraints; institutional ownership; and, retail order flow proxies. He constructs five models based on the progression of historical data availability for these variables as of 1926, 1950, 1976, 1996 and 2010. Using the specified data for every U.S. common stock as available during January 1926 through August 2025, he finds that: