Market participants, and academics, hotly dispute this question: how quickly does publicly available information get itself reflected in asset prices?
Stated that way, it is a huge question with a lot of subdivisions. Let’s focus on a portion of that for now: how quickly does information about a proposed merger or acquisition involving publicly traded corporations get itself reflected in the prices of the equity of the buyer and seller? The focus here is on the period between the announcement of a deal and the date of closing (or the date when it becomes clear the deal has for whatever reason failed, that there will be no closing).
The classic efficient-markets answer to that question is that the prices reflect the news either immediately or quickly enough to make the news reports useless as the basis for trades.
A recent investigation into the subject by Matthias Buehlmaier and Josef Zechner throws some doubt on that bit of classicism. Along the way, in their article “Media-Based Merger Arbitrage,” they reduce what might seem to be qualitative considerations to quantities, by estimating the probability that an announced merger will close from the length and the words used by media reports. I won’t try to detail their methods, but will cut beyond the chase to the capture.
Content versus Coverage
Buehlmaier, assistant professor of finance at Hong Kong University and Zechner, professor of finance at Vienna University of Economics and Business, maintain that information in financial media is not captured by same day price moves, or even the moves made within the first twelve trading days. If this is right, there is a lot of room for the capture of alpha through a trading strategy based on media content.
For their purposes, Buehlmaier and Zechner distinguish between media coverage and media content. Media coverage is the number of press articles on a given day that reference a given M&A deal. For a press article to count in this number, it must use both the name of the acquirer and that of the target within its first 100 words. Thus, given the “inverted pyramid” form of reporting, incidental mentions of a particular deal in the context of a story about something else are filtered out.
The authors perform certain calculations upon the resulting coverage number (I’ll spare you the specifics) in order to determine when the mentions reach a “surprising” level.
Separately, media content refers to both “the words used in a given press article, and … the article’s length.” The authors quantify content according to the implied opinion about whether a deal will actually close, which is of course the critical question for merger arbitrageurs. Thus, they score content along a scale from 0 to 1.0, the former indicating a successful merger is impossible, the latter indicating that it is certain.
A Winning Strategy
Surely an announcement in itself strongly suggests that probability of success should be well over 0! By their measures, working from a database that includes 130,589 press articles discussing 1,107 mergers, the average media content on the day of an announcement is 0.873. In the subsequent days it drops somewhat (if there are reasons for worrying about whether a closing will take place, they surface over time), to an average of 0.821.
A trading strategy based upon such numbers, expressing by way of one’s trading book the amount of confidence the media articles suggest about the closing of the deal, would be a winner. “[T]ime series tests show that the twelve trading days after an announcement date yield annualized risk-adjusted alphas of 18.5% for deals with a high media-implied probability of deal completion.”
Media coverage, that is the simple frequency of mentions, is much left helpful for an arb strategy, the authors write. They suggest that the ease with which such mentions might be manipulated by management of the firms involved may explain its disutility for arbs.