What is Sentiment Analysis?
Investors employ a variety of methods to analyse companies and the wider market, with the most popular methods being fundamental and technical analysis.
Those who employ sentiment analysis are concerned about the feelings surrounding the market or a stock. However, they do not base their decisions on their own opinions – rather they take cues from the beliefs of other investors in the market.
The sentiment of the market or stock adds a set of additional information that fundamental and technical analysis does not cover completely, which allows investors to make better investing decisions.
Sentiment analysis can be generally classified into two methods; the first seeks to predict market trends and reversals (market sentiment) while the second, enabled by recent data innovations, allows investors to predict individual stock returns (firm sentiment).
Sentiment analysis can be done systematically – defined as predictable returns and explainable by economics. Using large data sets, returns predictability can be achieved. A branch of financial economics, behavioural finance can well explain why sentiment analysis results in stock returns.
1. Market Sentiment
By understanding the sentiment of the broad stock market, investors can make a couple of investing decisions.
a. The first is to decide if investors should enter the equity market or allocate their funds to other assets. This is done by understanding if most of the investors in the market are feeling bullish (greedy) or bearish (fearful).
Several indicators can help to estimate the greed or fear levels in the market. One method is surveys, which are done via telephone interviews. The results are aggregated to produce a positive, neutral, or negative attitude regarding the market ahead. The University of Michigan Consumer Sentiment Index (US) and the BlackRock Investor Pulse Survey (international, includes Singapore) are two examples of such surveys.
b. The second is to decide if the market is trending – reaching a top or bottom (potentially a reversal) – or just having large or little volatility. Knowing such information allows the investor to calibrate his or her investment strategy and ride the waves like a pro.
One of the more popular indicators of such behaviour is the Volatility Index (VIX) published by the Chicago Board Options Exchange (CBOE). The index is a computation of the expected annualised change in the S&P 500 index over the next 30 days, based on options-based theory and current market data. The VIX was used extensively during the Global Financial Crisis of 2007 and the Euro Crisis of 2009 to judge investor sentiment.
However, investors should take note that market sentiment indicators like surveys, and the VIX are proxies of what the market is feeling and should be used on its own to trade.
2. Firm Sentiment
Due to advances in computation power and text analytics algorithms, investors can now measure what others are thinking regarding a specific company. Although it requires more advanced knowledge to handle data, textual sentiment on firms is getting more popular and easier to access.
The methods to extract sentiment from textual data are similar – investors have to employ text analytics to download, clean and classify words in an article to determine if the mood regarding a company is positive or negative. The differences lie in the data sources people use.
The first source of data is popular financial news platforms. Because many investors read articles on popular news sites like the Wall Street Journal or Financial Times, articles on these sites can influence mass swaths of investors. Researchers have even found that different authors on the same site have varying impact on stock performance1.
Another source of data is Twitter. Even though tweets are limited to 140 characters, the speed and volume of tweets coming from commentators can help to predict stock prices and volatility, especially tweets surrounding corporate announcements like earnings reports2.
The next method to understand investor sentiment regarding a firm is to look at internet search behaviour. Most use Google Trends and track search volume to judge if there is any abnormal attention on a company. One cannot judge if such attention is positive or negative, so it is less useful than the two previous methods, but it can allow investors to estimate if volatility is incoming.
Sentiment analysis is a lot of work, especially for those who focus on firm sentiment. But if done consistently, sentiment can be included with other forms of analysis to create mixed systems that are systematic and produce better results compared to just using evaluation methods individually.
1. Dougal, C., Engelberg, J., Garcia, D., & Parsons, C. A. (2012). Journalists and the stock market. The Review of Financial Studies, 25(3), 639-679.
2. Al Nasseri, A., Tucker, A., & de Cesare, S. (2014, October). Big data analysis of stocktwits to predict sentiments in the stock market. In International Conference on Discovery Science (pp. 13-24). Springer, Cham.