[This is another of a series of editorials by Executive Editor Andrew Karolyi at the Review of Financial Studies featuring forthcoming or recent papers at the journal. This editorial features “No News is News: Do Markets Underreact to Nothing?” by the University of Chicago Booth’s Stefano Giglio and Kelly Shue, lead article in Issue 27 (12) for December 2014. It was selected as an Editor’s Choice article on the Oxford University Press web site for RFS.]
The 1892 book, The Memoirs of Sherlock Holmes by Sir Arthur Conan Doyle, is a collection of short stories, one of which entitled “Silver Blaze” is a mystery about the disappearance of a famous racehorse the night before a race and the murder of the horse’s trainer. Sherlock Holmes solves the mystery by recognizing that no one that he interviewed in his investigation remarked that they had heard barking from the watchdog during the night in question. The fact that the dog did not bark when it was expected to while the horse was stolen leads Holmes to conclude the perpetrator was not a stranger to the dog and thus would cause it not to bark. This “negative fact” cracked the case.
The opening of Giglio and Shue’s paper makes clear that the tale of “the dog that did not bark” indeed inspired them in part to investigate a novel question: does the absence of news reports and the passage of time contain important information for markets? Their paper offers up many possible settings in which one could explore the instance of no news and how economic agents might react to its non-existence, but they offer up the context of mergers. As Shue writes to me, “the merger context is great for this analysis because we can easily quantify the information content of the passage of time: the time elapsed after a merger announcement without completion or withdrawal offers information about the probability a merger will be completed.” The authors build a simple hazard rate function of the likelihood of completion in the next week conditional on its not having been completed or withdrawn to date. The implied hazard rate function forms a hump shape in a consistent matter across a large sample of U.S. merger deals over a 35-year horizon. The most intriguing part of the paper comes when the authors show that the implied hazard rates can predict weekly returns: the higher the probability of completion, the higher is also the average return.
The study takes yet one more turn in the mystery, making Sir Arthur very proud. A kind of mispricing is revealed. With the passage of time after merger announcement with hazard rates rising, the market seems to underestimate the probability of merger completion and positive surprises arise on average. As even more time passes and hazard rates start to fall, the market overestimates the probability of merger completion and negative surprises ensue. The intensity of this under- and over-reaction is greatest for illiquid stocks which may imply some kind of limits to arbitrage.
When I asked the authors to rationalize the big takeaway from their study, they offered that this evidence of mispricing to the passage of time is “suggestive of a more general phenomenon…it is less exciting than the new stories typically covered by the media.” They added that “underreaction to new news can potentially exacerbate asymmetric information problems in other contexts such as between voters and politicians, between managers and employees, or investors and insiders, where arbitrage is even more difficult.”