Media content, sentiment and emerging market futures returns

ABSTRACT

The paper attempts to capture the sentiment derived from routine financial news and outlines the impact of the media content on the main index futures contracts of Hong Kong and Singapore. News factors are generated from routine financial news, but pessimistic market sentiment factors are more prevalent. High pessimistic news factors predict lower returns for the same day, and the returns start to reverse two days after the news for Hong Kong and Singapore markets. The finding is consistent with the sentiment theory. There is a significant reversal in returns, and the reversal in returns offsets the initial changes entirely. The bad news factor does not seem to work as a proxy for trading costs. The sub-sample results are similar to the whole sample. We also find the significant impact of U.S. news sentiment on news factors, futures return, and open interests in both markets. Trading strategies based on bad news factors generate economically significant returns when trading costs are considered.

KEYWORDS:

• Investor sentiment
• news
• media
• futures market

JEL CODES:

• G13
• G14
• G15

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ One of the reasons why U.S. stocks market appears to have a disproportionate effect on markets of other countries is the United States uses the English language, which has emerged as a world language. It is much easier for foreign reporter, who invariability knows English, to respond to stories from the United States or the United Kingdom than to stories from Germany or Brazil. Producing news stories is a business with tight deadlines. And it requires fast actions. A lot of reporters have the ability to pick up a story from another country in English, and turn it into a local story in pinch. (Shiller, 2005, p. 104).

² The initial margin is HKD90,450 and SGD4,750 for Hang Seng Index Futures (HSIF) and Morgan Stanley Singapore Free Index Futures (SiMSCIF) respectively. Final settlement may incur huge amount of cost because contract multiplier is HKD50 per index point for SHIF and SGD200 for SiMSCIF. Top ten HKEX participants hold 87.68% long open interest of Hang Seng Index futures from 30^th March to 3rd April 2009. During the same period, it was also found that the top ten HKEX participants hold 78.37% short open interest for the same contract.

³ This is to transform the trading volume and open interest into a stationary time series. Campbell, Grossman, and Wang (1993) and Fung and Patterson (1999) employ the same method, but 100-day backward moving average are used. Fung and Patterson (1999) have also tried with 20-day backward moving average and yield similar results.

⁴ We calculate the returns as follows: 24-hour return as $\left(log\left(\frac{P_{settlement,t}}{P_{settlement,t-1}}\right)\right)$, overnight returns as $\left(log\left(\frac{P_{open,t}}{P_{settlement,t-1}}\right)\right)$ and day-time returns as $\mathrm{l}\mathrm{o}\mathrm{g}\left(\frac{P_{settlement,t}}{P_{open,t}}\right)$.

⁵ The proxy for volatility is calculated as follows. Firstly, the variable R_t is regressed on 12 lags of R_t to obtain a residual. We use Bayes information criterion (BIC) and the Schwarz information criterion (SIC) for selecting the lag length. The residual is then squared, and a past 60-day moving average is subtracted from the squared residual. Tetlock (2007) defines the trend as 60-day moving average; Ciner (2006) defines the trend as 200-day moving average and Bessembinder and Seguin (1992) define the trend as 100-day moving average. These studies conclude that the results are robust to the number of days used to calculate the trend.

⁶ Yen et al. (2001) finds a consistent up-moving trend from 15 days before Chinese New Year and lasts up to 15 days after that.

⁷ For Hong Kong, the Hang Seng Index is the major performance benchmark for the Hong Kong equity market since 1969. It is weighted by market capitalization of 33 constituent stocks. Hong Kong Futures Exchange introduced the Hang Seng Index Futures (HSIF) contracts in May 1986. The HSIF recorded total trading volume of 21,716,508 contracts in the year 2008. In the same year, the range of average daily trading volume is from 74,311 to 111,456 contacts. (source: HKEx Fact Book 2008, http://www.hkex.com.hk/data/factbook/2008/fb2008.htm).

⁸ The SiMSCI is constructed by the Morgan Stanley Capital International. The index consists of 36 stocks from five sectors, namely financials, industrials, telecommunications services, consumer discretionary and consumer staples. Futures contracts of these indexes are electronically traded on the Singapore Exchange Derivatives Trading (SGX-DT).

⁹ HSIF and SiMSCIF switch over on last trading day of the nearby contract month, i.e. the second last business day of the month.

¹⁰ The 9 word categories are: first, Positive (related to positive outlook); second, Strong (indicating strength); third, Active (regarding active orientation); fourth, Rise (indicating movements); fifth, Negative (associated with negative outlook); sixth, Weak (implying weakness); seventh, Passive (regarding passive orientation); eighth, Fail (indicating goal has not been achieved) and ninth, Fall (indicating movements).

¹¹ For robustness check, we also use trading volume measured as the number of contracts traded in section 7..

¹² In the context of equity index in the spot market, Tetlock (2007) finds Pessimism has negative impact on the next day Dow Jones’ returns.

¹³ For robustness, we also estimate the impact of the absolute change in news factor on the percentage change of de-trended log of trading volume, which is analysed in Section 7.

¹⁴ Appendix C provides the examples of these strategies considering the price of long and short prices, and prices at maturity. These example shows that the overall profit is the same for the two strategies regardless of the price at maturity. Similarly, the total loss is also same if the index futures price goes against the prediction based on the news factor.

¹⁵ This is based on news lease by the HKEX, dated 16^th August 2001. The round-trip trading cost includes brokerage commissions, exchange trading fees and stamp duty. http://www.hkex.com.hk/eng/newsconsul/hkexnews/2001/010816news.htm, Accessed on 26th April 2010.

¹⁶ The index is available from 1980. Shapiro, Sudhof, and Wilson (2020) find that the daily news sentiment index can predict the survey-based measures of consumer sentiment. The sentiment index in the days leading up releases of the Michigan Consumer Sentiment Index and the Conference Board’s Consumer Confidence Index is strongly predictive of those releases. Moreover, any positive innovations to sentiment increase consumption, output, and the real fed funds rate, and temporarily decrease inflation.

¹⁷ For brevity, we do not report the results with tables.

¹⁸ In order to limit the number of tables, we exclude the tables of sub-sample results. These are available upon request..