Sentiment analysis has been a tech buzzword which entailed the popularity of social media.
Just as the child bites at almost everything as soon as its milk teeth grow, everyone with a social media account seemed to have an opinion about everything. And companies were trying to exploit this tremendous amount of data by analyzing public opinion.
I have also been involved in such an effort though I now admit that our sentiment engine was not good enough.
But what worries me is another buzzword predictive analytics. And its dangerous application in stock markets.
Several startup companies are trying to predict stock price based on public sentiment. The practice of predicting stock prices itself is a questionable business model; seems they are trying to justify the prediction using this new methodology.
Information that triggers price movement is private to only a few who take advantage as early movers. Hence the price trend already starts way before the news is published in mainstream media.
Look at TATA STEEL.
The general sentiment in beginning of Mar 16 on Indian steel industry was extremely negative. China had produced more steel in one year than the Brits had done in last 80. However, we see a 30% increase in TATA STEEL price in the same month. Only at the end of the month, the company made public of its decision to sell Chorus. Under the circumstances it would reduce debt, hence the optimism about the company. Selected few knew that their CFO was making several trips to London and had taken advantage of this fact.
Another example of a debt ridden company - GMR infrastructure.
Look at the sharp price rise on 13 Jun 2016. It was after four days the general public learnt about its progress to sell off stake in a profit making project.
(as of writing this article, the company had already denied it as rumors !! ) People interested may track the story further to see how such things wipe off hard earned money of retail investors.
Hence we can conclude the following.
One can look at the currently emerging price trends and predict the future sentiment but not the other way around.
Just as the child bites at almost everything as soon as its milk teeth grow, everyone with a social media account seemed to have an opinion about everything. And companies were trying to exploit this tremendous amount of data by analyzing public opinion.
I have also been involved in such an effort though I now admit that our sentiment engine was not good enough.
But what worries me is another buzzword predictive analytics. And its dangerous application in stock markets.
Several startup companies are trying to predict stock price based on public sentiment. The practice of predicting stock prices itself is a questionable business model; seems they are trying to justify the prediction using this new methodology.
Information that triggers price movement is private to only a few who take advantage as early movers. Hence the price trend already starts way before the news is published in mainstream media.
Look at TATA STEEL.
The general sentiment in beginning of Mar 16 on Indian steel industry was extremely negative. China had produced more steel in one year than the Brits had done in last 80. However, we see a 30% increase in TATA STEEL price in the same month. Only at the end of the month, the company made public of its decision to sell Chorus. Under the circumstances it would reduce debt, hence the optimism about the company. Selected few knew that their CFO was making several trips to London and had taken advantage of this fact.
Another example of a debt ridden company - GMR infrastructure.
Look at the sharp price rise on 13 Jun 2016. It was after four days the general public learnt about its progress to sell off stake in a profit making project.
(as of writing this article, the company had already denied it as rumors !! ) People interested may track the story further to see how such things wipe off hard earned money of retail investors.
Hence we can conclude the following.
One can look at the currently emerging price trends and predict the future sentiment but not the other way around.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.