Foreign Exchange Volatility, Media Coverage, and the Mixture of Distributions Hypothesis: Evidence from the Chinese Renminbi Currency

Abstract

The stability of Chinese renminbi (CNY) against the US dollar (USD) has been the recent focus of economic and political discussion in the media, especially in light of the growing political pressure exerted by the US government to let the CNY/USD appreciate. This paper contributes to the discussion of the CNY/USD stability by examining its volatility dynamics since the floating regime is established. In particular, we analyze the fluctuations of the CNY/USD from the perspective of the mixture of distributions hypothesis (MDH), which argues that the variability of financial assets at a given interval is proportional to the rate of public information arrivals. By using the well-known news database FACTIVA, which offers a comprehensive collection of the world's top media outlets and publications (such as The Wall Street Journal and The New York Times) in Chinese and English languages, we document the statistical significance of the impact of information arrivals on the conditional volatility of the CNY/USD. More specifically, information arrivals are measured by the number of news items that appear in the various major media sources for each day from July 21 2005 (which marks the first day of the floating regime). The media sources comprise not only the top ten English-language US newspapers (by circulation numbers) but also the twelve major Chinese-language media publications in the Asia-Pacific region (including five from mainland China, two from Hong Kong SAR, and three from Taiwan). These Chinese publications are selected on the basis of wide regional and global readership. The relevant news items from all the media sources can be broadly classified into two categories: macroeconomic news containing relevant information on the CNY exchange rate and political news on the Sino-US relations. We adopt two main approaches to analyze the link between public news and the CNY/USD volatility. First, by applying the Generalized Autoregressive Conditional Heteroskedasticity (GARCH) framework that incorporates asymmetric effects and conditional leptokurtic distributions to the daily spot CNY/USD data for the period from July 21 2005 - December 31 2010, we observe that there is a significantly positive impact of all the daily public news items on the conditional volatility of CNY/USD. This result is robust across alternative model specifications and distributional assumptions. Amongst all the different models, the Exponential GARCH-in-Mean (EGARCH-M) with conditional Student-t distribution works best for the daily data. In particular, compared with the EGARCH-M model with no news effects, the volatility persistence in the EGARCH-M model with news is reduced. As a further robustness check, our analysis is also extended to the daily non-deliverable forward (NDF) rates for various maturities. Consistent with the results for the spot rates, public news arrivals have a positive impact on the conditional volatility of the NDFs. Our second approach is to use the framework of Bauwens, Rime, and Sucarrat (2006) by computing realized volatility measures for weekly CNY/USD rates. Consistent with the results obtained for the daily data, public news has a positive impact on CNY/USD volatility. Range-based volatility measures are also used and they confirm the positive link between exchange rate volatility and public news. In general, our results suggest that the intensity of media coverage is a significant determinant of the volatility dynamics of the CNY/USD exchange rate.