Can you trade on rumors? One possible model

In Tradestream, I spent a whole chapter looking under the covers of investing/trading strategies that focus on rumors.  Greater sentiment analysis (like the hubbub that erupted after Prof. Bollen published a paper on using twitter to predict stock market swings) is in early days but at its core is a desire to use news/chatter to better gauge future stock moves.  Bold and audacious, but not nearly there yet.

Rumors: A Model

In the book, I developed a model upon the one Cass Sunstein (co-author of Nudge and just a prolific writer/thinker) used in his most recent book, On Rumors: How Falsehoods Spread, Why we Believe Them, What Can Be Done.

Rumor transmission often involved the rational processing of information, in a way that leads people quite sensibly in light of their existing knowledge, to believe and spread falsehoods.  This problem is especially acute on the Internet -- Cass Sunstein, "She Said What?  He Did That? Believing False Rumors," Harvard Law School Public Law Working Paper No. 08-56 (November 2008), 2

Sunstein describes a useful framework with which to understand how rumors get started and how they get propagated -- influencing decision making. Sunstein describes the various actors in the social transmission of false information.  While he focuses on rumormongering, I try to apply this framework to investing.

1. Propagators:
   1. self-interested, varying degrees: they may own a stock and work to discredit those who don't like it or are short
   2. altruistic: sincerely interested in promoting some type of cause -- these guys don't even realize that they are spreading falsehoods
2. Priors: success or failure of rumors depends on how closely they approximate the prior beliefs of those who hear them
   1. motivations: people don't enjoy hearing bad things about ideas/people close to them and conversely, they are more open to receiving false info about something they dislike
   2. beliefs: Sunstein says that people who have strong prior beliefs usually do so because of what they know and therefore, require a lot of supporting information to upseat those beliefs
3. Cascades: the mechanisms of rumor transmission, why/how/when people accept/reject a rumor is intimately connected to how the information affects their personal desires
   1. informational: groups of investors are led to accept a thesis in spite of individuals' private info.  Think of all the hating that goes on on Yahoo Message Boards.
   2. reputational: people can be led to believe things in conflict with their priors but do so to curry favor with others.  This is equivalent to a fund manager on CNBC pumping his portfolio -- as an expert -- his status and street cred influence others' beliefs (whether correct or not)

So, we have to narrow our focus down to why stocks move they way they do when unsubstantiated news -- rumors -- are floated.

Rumors and Preannouncement Trading

I chose to focus on rumors surrounding M&A announcements.  Many times, the Wall Street Journal will publish stories on unsubstantiated mergers and acquisitions.  Target stocks will jump and acquirer stocks drop.  That said, though, many of these rumored M&As fail to consummate.

"While sellers lose money when a rumor precedes an actual announcement, in most cases rumors fail to materialize into public announcements." Rumors and Pre-Announcement Trading: Why Sell Target Stocks Before Acquisition Announcements?" (Gao, Oler)

Given the research of Gao and Oler:

On average, stock prices of rumored firms drift down to their pre-rumor level over a 70-day period after the initial price jump when a rumor is published and that only 12% of rumored takeovers materialize into actual announcements within 70 days.

So, really, Tradestreaming would be all about finding the right side of this strategy -- where the numbers, data and probability is with the investor.  That would mean taking the other side of the trade.

The Antitakeover Strategy

1. Research WSJ for reported but unsubstantiated M&A
2. Remove all mega cap firms (<$20B)
3. Short a basket of rumored acquisition targets and hold 70 days after the rumor first appeared.  You can hedge by going long the market if you like.
4. Strategy performs even better during periods of increased M&A activity

   

Performance

The researchers found that this strategy would put up 4.2% in abnormal returns -- when you further restrict the strategy to hot M&A years, profits go up to 12.7%.