News sentiment is able to capture the changing patterns in various sectors of the economy and high degree of co-movement is seen with key macroeconomic indicators, according to an article in RBI’s latest monthly bulletin.

Empirical analysis reveals that media sentiment indices have a high degree of co-movement with economic indicators and optimistic sentiment periods coincide with higher growth phases and vice versa, opined the authors of the article “When a News Story is More Than Just Text: Evidence from Indian Economy”.

Effective forecasting

Further, the sentiment indices possess predictive ability in terms of providing directional change in underlying economic variable in nowcasting and short-term forecasting

The authors examined the informative value contained in the news sentiment through logistic and tree based modelling, and found that it predicted the directional change in the corresponding target economic indicator, well in advance.

A good tracking ability of sentiment indices coupled with the fact that news sentiment is available in near real time, prior to the release of the actual hard data, makes the sentiment more useful, RBI officials (Geetha Giddi, Shweta Kumari and Sasanka Sekhar Maiti) said.

Through their deep dive exercise, the authors collated the scattered information contained in news, which can reflect media sentiment about economy on an ongoing basis and presented the same in a systematic manner.

The authors extracted economic signals from online print media, built on natural language processing techniques and fine-grained aspect-based sentiment analysis.

“The natural design of news articles - unstructured, noisy, complex and influenced by a variety of factors - present challenges and requires utmost effort in order to achieve efficiency.

“Insights present in the news text, available at high frequency, are assessed to capture the tangible effects for macroeconomic and business cycle indicators relevant from central banking perspective,” the authors said.

News media offers perspective on concurrent topics encompassing multiple sectors of the economy, per the article.

“Globally, literature studies suggest that news-based sentiments possess valuable economic signals and are increasingly being utilised in nowcasting and forecasting framework,” it added.

As the sentiment indicators can be constructed on high frequency basis, prior to the release of official statistics and traditional indicators, it is worthwhile to track them regularly for early signals on economic conditions, the RBI officials said.

Further extensions can be carried out by incorporating other commonly used measures of confidence, sentiment and expectations of market agents as available through business tendency surveys, and estimating wholesome empirical models aiming at higher prediction accuracy.

comment COMMENT NOW