Using Twitter and Information Cascades to Predict the Market
This NY Daily News article discusses the prospect of accessing the social network Twitter as an informational resource to predict the future three days in the market to about an approximately 87% accuracy. The article goes in depth into how monitoring Twitter feeds as an individual investor would not be a wise choice; if major professional traders were to employ the same tactics with their more sophisticated systems, they would be able to more efficiently and proficiently trade with algorithms based on Twitter. The author discusses how although the market may be predicted to an 87% accuracy for up to three days, it is based on mainly emotional reactions, and monitoring Twitter is not a great way to invest or trade in the market – the author implores the readers to invest/sell based on the fundamentals of companies instead of emotional responses of Twitter users.
This article touches on several topics covered in our course work including trading networks, information cascades, and social networking. The author’s argument stems from concepts that can be compared to those in information cascades. In an information cascade, people observe the actions of others and make similar decisions as them. For example, as the author argues, if someone notices that the price of a company’s stock is going down, they get disheartened or nervous that the company is going down fast and they are losing their money – therefore it would probably be ludicrous to keep your stock…SELL SELL SELL! If a lot of people are selling their stock…maybe you should also sell your stock? Well, based on an information cascade, yes, this is true: sell. However, as discussed in the article, what you should do instead is maintain some discipline and analyze the fundamentals of the company in question. If you see that the company’s stock is going down, but their earnings estimates are still solid and their PE is lower than the growth rate, then you should actually be purchasing stock, not selling. If an investor were to follow the emotional information from Twitter, yes, they would be able to predict the market for three days at say . . . an 87% accuracy; but you would not be making the correct decision unless the company’s fundamentals were also decreasing (i.e. NFLX).
This example excellently illustrates the dangers in information cascades; dumping your shares in a company based on an information cascade will hurt a company’s stock in the short run, but in the long run will ultimately hurt the investor the most assuming the company’s earnings are in the opposite direction of the cascade’s predicted outcome.
Thank you. Have done a nice article sharing. I follow your site constantly …