Advanced

Language, News and Volatility

Byström, Hans LU (2014) In Working Paper / Department of Economics, School of Economics and Management, Lund University
Abstract
I use Google News TM to study the relation between news volumes and stock market volatilities. More than nine million stock market-related news stories in English and (Mandarin) Chinese are collected and the dynamics of the news volume and the stock market volatility is compared in both the Anglophone world and the Sinophone world. I find that the stock market volatility and the number of publicly available global news stories are strongly linked to each other in both languages. Contemporaneous correlations between news and volatility are positive and highly significant, and regressions tell us that the directional link between news and volatility rather is from news to volatility than vice versa. In out-of-sample evaluations of volatility... (More)
I use Google News TM to study the relation between news volumes and stock market volatilities. More than nine million stock market-related news stories in English and (Mandarin) Chinese are collected and the dynamics of the news volume and the stock market volatility is compared in both the Anglophone world and the Sinophone world. I find that the stock market volatility and the number of publicly available global news stories are strongly linked to each other in both languages. Contemporaneous correlations between news and volatility are positive and highly significant, and regressions tell us that the directional link between news and volatility rather is from news to volatility than vice versa. In out-of-sample evaluations of volatility forecasts I find news volumes to improve forecasts, regardless of language. The relationship between news and volatility is weakest in mainland China and a possible reason for this is that Chinese retail investors do not read (traditional) news, neither in Chinese nor in English. The results suggest that news could be used in volatility-related financial applications such as GARCH-models or VIX-like fear indexes. (Less)
Please use this url to cite or link to this publication:
author
organization
publishing date
type
Working Paper
publication status
published
subject
keywords
news aggregator, news, language, volatility, stock market, Chinese, Mandarin, GARCH, VIX
in
Working Paper / Department of Economics, School of Economics and Management, Lund University
issue
41
pages
43 pages
publisher
Department of Economics, Lund Universtiy
language
English
LU publication?
yes
id
59cc6906-da60-4acb-a91e-a767186b16fd (old id 4865070)
alternative location
http://swopec.hhs.se/lunewp/abs/lunewp2014_041.htm
date added to LUP
2014-12-17 15:53:18
date last changed
2016-09-13 16:18:50
@misc{59cc6906-da60-4acb-a91e-a767186b16fd,
  abstract     = {I use Google News TM to study the relation between news volumes and stock market volatilities. More than nine million stock market-related news stories in English and (Mandarin) Chinese are collected and the dynamics of the news volume and the stock market volatility is compared in both the Anglophone world and the Sinophone world. I find that the stock market volatility and the number of publicly available global news stories are strongly linked to each other in both languages. Contemporaneous correlations between news and volatility are positive and highly significant, and regressions tell us that the directional link between news and volatility rather is from news to volatility than vice versa. In out-of-sample evaluations of volatility forecasts I find news volumes to improve forecasts, regardless of language. The relationship between news and volatility is weakest in mainland China and a possible reason for this is that Chinese retail investors do not read (traditional) news, neither in Chinese nor in English. The results suggest that news could be used in volatility-related financial applications such as GARCH-models or VIX-like fear indexes.},
  author       = {Byström, Hans},
  keyword      = {news aggregator,news,language,volatility,stock market,Chinese,Mandarin,GARCH,VIX},
  language     = {eng},
  note         = {Working Paper},
  number       = {41},
  pages        = {43},
  publisher    = {Department of Economics, Lund Universtiy},
  series       = {Working Paper / Department of Economics, School of Economics and Management, Lund University},
  title        = {Language, News and Volatility},
  year         = {2014},
}