New stock-price predictor: Amazon reviews

This article was originally published on this site

Investors might want to start paying attention to a different kind of five-star rating.

A new study suggests that AMZN, +2.16% product ratings can be used to predict a company’s stock returns. Jiekun Huang, an assistant professor of finance at the University of Illinois at Urbana-Champaign, analyzed nearly six million Amazon reviews from 2004 to 2015, and found a positive relationship between customer ratings and a company’s subsequent stock returns.

“I’m a bit surprised about the results,” says Huang, whose previous research focused on the behavior of institutional investors. “In aggregate, customer reviews may provide useful information.”

The idea is that the cumulative power of consumer opinion, in the form of positive reviews, conveys information about consumer goods, Huang says. When consumers view certain products favorably, it indicates higher product quality and value, which predicts higher cash flows and, in turn, higher stock prices. The study wasn’t an attempt to test whether consumer opinions cause stock prices to change, Huang says; rather, it was a test of whether consumer opinions contain novel information that hasn’t already been incorporated into a firm’s share price.

Range of industries

Huang’s paper, which hasn’t yet been published, examined reviews of 133,750 products manufactured by 231 public companies in industries ranging from business equipment to nondurable goods. Companies studied included Mattel Inc., MAT, +0.43% Coach Inc., COH, -0.70% Estée Lauder EL, -0.32% and SanDisk Corp. (which was bought this year by Western DigitalCorp.). WDC, +0.49%

To maintain consistency in his sample, Huang made sure each product had at least 10 reviews in a month and then averaged the star ratings for each product on a month-to-month basis.

The paper found, among other things, that stocks associated with higher-rated products in one month outperformed those with lower ratings by around 0.8 percentage point in the next month. Extrapolated over a yearly period, that’s a total of about 10 percentage points.

Of course, not everyone believes consumer reviews contain value-relevant information. They point out that not all customers tell the truth in their product reviews, and some don’t have the expertise to review products adequately.

Still, academics who have looked at Huang’s paper think his findings are generally sound, though some say more research needs to be done to tie them up. “It makes intuitive sense, but it’s not yet clear that it’s not capturing something else,” says Avanidhar Subrahmanyam, a professor of finance at the UCLA Anderson School of Management.

For one, Subrahmanyam says, the study didn’t control for variables such as advertising and a company’s profit margins, which also can contribute to an increase in returns. It also isn’t clear, Subrahmanyam says, that Huang can make reliable inferences from his data set since his study, on average, looked at just 83 stocks each month.

Huang agrees that his research could be furthered, and he is in the process of revising his paper to control for certain variables. He also says analyzing the text of Amazon reviews, rather than just the star ratings, could provide more worthwhile information.

By chance

Huang’s paper contributes to a body of research that focuses on the investment value of “serendipitous information,” or information that investors come across by chance, according to Paul Tetlock, a professor of finance and economics at Columbia Business School.

Serendipitous information, he says, could include anything from Amazon reviews to trends in Google searches for products to satellite images of retailers’ parking lots.

“Without access to all possible sources, it’s hard to pinpoint the original source of the information,” Tetlock says. “But each of these sources likely embodies meaningful business trends, which could be reflected in stock prices.”

Matthew Kassel is a writer in New York. He can be reached at

The article “New stock-price predictor: Amazon reviews” first appeared on

More from MarketWatch