Michael Lewis called them “Flash Boysopen in a new window,” who use algorithms to predict and profit from significant moves made by slow-moving funds. These algorithms detect when mutual funds are about to purchase large stock quantities and can direct the system to start buying shares immediately. These algorithmic traders then profit from a short but predictable rise in stock price when mutual funds move.
Automated trading is said to make the market more efficient on two levels. It speeds up the price discovery process because the algorithms can infer from previous trades and quotes what smart money people already know. It also leads to increased trading volume, which could reduce the costs of individual transactions.
A study has found strong evidence that automated traders also drive out traders who gain their edge through fundamental research. This crowding-out leads to less fundamental analysis of small-company stock and less information on the market. Slowing down speed demons may benefit the economy and the markets, says Charles Lee, professor of accounting at Stanford Graduate School of Business. He co-authored the study with Edward Wattsopens Window, a former Stanford GSB Ph.D. student, now a Yale University faculty member.
Tick Talk
Lee and Watts examine what happens when automated traders encounter a temporary speedbump in their study. The Securities and Exchange Commission’s two-year pilot study on the effects of “tick sizes” (the minimum increment that a stock price can rise or fall) was the source of the vast amount of data they used.
Before 2005, stock prices were often quoted in fractions of dollars. The minimum tick size was usually one-sixteenth or 6.5 cents or more. Since 2005, the decimalization of stock prices has reduced the tick size to one cent. This change has boosted automated trading because it is much cheaper to trade stocks quickly when the price increments are minor. In 2016, the SEC experimented to determine if a smaller tick size is good or bad for stocks of small companies. In 2016, the SEC tested whether a smaller tick size was good or bad for small-company stocks.
Lee and Watts confirmed what many experts predicted: the larger tick size caused a significant drop in algorithmic trades. The automated trading fell by nearly 11% for stocks with higher tick sizes. They also found that fundamental research was increasing: searches for these companies’ filings at the SEC website soared, particularly in the weeks leading up to their earnings announcements.
Stock prices for companies with less automated trading began to reflect earnings news more accurately. Lee and Watts used 60-day returns to measure how much those stock prices anticipated the earnings news for each quarter. They reasoned that if increased fundamental activity in firms with larger tick sizes led to more informed stock prices, their pre-announcement returns should better predict the upcoming earnings news. They found that the pre-announcement returns of firms with tick sizes increased to five cents and contained more information on the upcoming earnings.
Elephant Gnats
Why would algorithmic trading lead to a less informed market? Lee believes automated trading drives out traders who gain their edge through fundamental research. Imagine that you are a portfolio manager trying to gain an advantage using fundamental analysis. Your insights will be much less valuable if a trading system can predict your trades before you have a chance.
Trading algorithms allow for tactics like “back-running,” where a system watches for signs that an institution is about to buy or sell stock and then moves ahead. The large fund managers know they can’t buy huge blocks of stores simultaneously. They, therefore, divide their orders into smaller ones. They often have to purchase orders that are harder to execute first thing in the morning. An algorithmic trading system quickly recognizes those early purchases and starts buying the stock immediately in anticipation of the larger fund’s significant assets temporarily driving up its price. Lee compares algorithmic traders to gnats in an elephant.
Lee says that automated trading is a form of poaching. The algorithm does not care about the fundamentals or analysis but can extract insights from someone else’s trading behavior. There are many algorithmic traders, and they can discourage other traders from investing in original research. This is especially true for smaller companies that have lower trading volumes.
Lee warns that this study was designed to focus on smaller companies, and the effects of tick size could differ for companies with high valuations. A change in minimum tick sizes may not have the same impact on giant companies with millions of investors.
The pilot project has reduced automated trading, but at least in the case of small businesses. This seems to have improved the market’s sensitivity. Lee claims that the pilot project had a “cleansing effect” on pricing efficiency. The elephants returned after the gnats left.