Berkeley - Internet stock prices generally rise before earnings announcements and fall afterwards, regardless of the earnings and revenues reported, according to a new study by three accounting professors at the University of California, Berkeley's Haas School of Business.
This first-ever academic study of Internet stock fluctuations surrounding earnings announcements found a predictable price pattern with a 10-day return of more than 11 percent from January 1998 to August 2000.
The study conducted by professors Brett Trueman, M.H. Franco Wong and Xiao-Jun Zhang found a general run-up in prices in the days prior to an Internet firm's earnings announcement and extending through the market opening on the day subsequent to the release. This run-up is followed by a price reversal lasting several days. The price pattern persisted even during and after the crash of the technology stock market in October-November, 2000.
Purchasing shares five days before an Internet firm's earnings announcement and holding them through the market opening on the day after the announcement generated a return of about 5 percent, on average. Short-selling at the time, followed by short-covering five days later, earned on average another 6 percent.
Short-selling refers to borrowing shares of stock from a broker or dealer and selling them on the open market. Short-covering refers to repurchasing a stock previously sold short.
"We were surprised by the size of these returns and by the fact that they occurred quarter after quarter...whether the market was going up or down," said Zhang.
The researchers also found that, in general, neither the announced earnings nor the announced revenues are related to the pre-announcement price run-up or post-announcement decline.
"This rules out the possibility that these returns are due to news leakage in advance of the earnings announcement," according to Wong.
Rather, the evidence points to excessive demand for shares in advance of the earnings announcement, a situation that temporarily drives up prices. Trading volume, for example, was found to be much higher than normal in the few days before the earnings announcement and returned to more normal levels shortly after the release.
"It appears from this evidence that small investors were buying just before Internet firms' earnings announcements in the hopes of good news, which would drive the prices up even further," said Trueman. Given the subsequent stock price declines, this is clearly an unprofitable strategy, he said.
Overall, the researchers' results suggest some degree of irrationality in the pricing of Internet stocks around earnings announcements.
The researchers emphasize that their findings hold on average, and that not every individual stock exhibits the same pattern around earnings announcements.
"Internet stocks are very volatile. Trying to exploit our findings by taking a position in just one or two stocks is very risky. Diversifying across a large number of Internet stocks reduces that risk," said Zhang.
The study sample included roughly 400 publicly traded Internet firms. Similar, but smaller, stock price patterns were found for a sample of almost 1,000 non-Internet high-tech firms. These patterns were not present for a selected sample of more than 300 non-technology firms.