Short Seller Claims Chinese Ed-Tech Firm Used Bots to Inflate Its Market Value

Staff Writer

A U.S.-based short seller claims that Chinese ed-tech company GSX Techedu is engaging in fraud, saying in a recently published report that it is highly confident that approximately 73 percent of GSX’s reported users are actually bots.

GSX is a K-12 after-school tutoring service provider in China, offering courses covering all primary and secondary grades, as well as professional and interest courses.

The Chinese company disputed the claims, made by the investment firm, Muddy Waters Research, saying the short seller “lacks a basic understanding” of GSX’s operations.

As recently as May 18, the date that short seller Muddy Waters Research’s report was published, GSX’s stock price was $35.43 per share.

Muddy Waters based its conclusions on what it says are 463,217 sign-on records for 54,065 unique users of GSX. The short seller identified three “bot patterns” that Muddy Waters believes accounted for 73 percent of the study’s unique users.

It’s possible that about 81 percent of the platform’s users are bots, according to Muddy Waters.

The research firm says it identified about 53 percent of the unique users in its sample as bots because they are “precise joiners,” or users who join a class at the same time, to the second, on the same day of at least two different weeks. The short seller said the probability of such a situation happening two or more times for a single user in a single course is extremely low.

Further, Muddy Waters says that it counted thousands of “burst joiners,” referring to large groups of users who join at the same second, often in time periods otherwise marked by little to no joining activity.

“This anomaly strikes us as similar to seeing 10 subway trains pass by in an hour, with nine of them completely empty and one completely full,” the short seller’s report says. “Real life just does not work this way.”

Short sellers are investors who target companies they believe are overvalued and then bet on the decline of those companies’ stock at a future date. As a result, they stand to gain if the stock price plummets. Muddy Waters has taken aim at numerous companies in the past. including publicly traded firms based in China.

In a response refuting Muddy Waters’ allegations, GSX Techedu said burst and precise joiners are brought about naturally by the structure of GSX’s business operation, which splits a large class with one instructor into smaller groups with multiple tutors.

According to GSX, which is traded on the New York Stock Exchange, burst joiners are caused when classes transition from tutors to instructors, a typical course procedure.

“Typically, a tutor offers a prep session before a paid class to warm up students with games and quizzes,” the GSX statement says. “When the formal class starts, the tutor will transition all his or her students from the small groups to the instructor’s large class, and the instructor takes over.”

If a tutor fails to take over in time, the system “will automatically switch classes for students, usually around the class start times, thus causing large groups of users to join classes at the same time,” GSX Techedu said.

Muddy Waters’ May 18 report follows an April 14 report by short seller Citron Research, which argued that up to 70 percent of GSX’s revenues are fabricated.

GSX also denied the allegations in Citron’s report.

Muddy Waters Research also claimed that an unidentified former GSX manager corroborated the short seller’s observations about “fake user patterns” associated with the Chinese firm, saying that the company’s bot operations started in 2015.

Follow EdWeek Market Brief on Twitter @EdMarketBrief or connect with us on LinkedIn.


See also:

One thought on “Short Seller Claims Chinese Ed-Tech Firm Used Bots to Inflate Its Market Value

  1. Citron Research calls GSX “the most blatant Chinese stock fraud since 2011” and asks that trading of its shares be halted and for an immediate internal investigation in a report published on Tuesday.

Leave a Reply