News Sentiment Investment Strategies in APAC - A Case Study

RavenPack | May 21, 2020

Trading strategies based on Asia Pacific news sentiment produced excess returns for up to a month ahead

Apply News Sentiment to Investment Strategies in Asia-Pacific Markets: Methodology

Does news sentiment data work in the Asia Pacific region, and if so, what are the results from using it as the basis for a trading strategy? These were two of many questions the RavenPack Data Science team tried to answer in a recent white paper “Applying News Sentiment to Investment Strategies in Asia-Pacific Markets”.

To find the answers they backtested sentiment-based equity strategies on portfolios containing stocks from three broad regions: Australia & New Zealand, Japan, and Asia ex-Japan; as well as a pan-Asia-Pacific portfolio, that combined all the individual markets into one.

Buy and Sell signals were generated using scores from a daily company sentiment indicator built from RavenPack’s Event Sentiment Score (ESS) over a 24-hour lookback window.

News Sentiment Provides Alpha in APAC

The overall conclusion appeared to be that yes, news sentiment-based strategies did work in APAC, especially for short-to-medium term holding patterns, with the region as a whole producing Information Ratios close to 4.0, and delivering greater than 20% Annualized Returns.

Performance APAC Regional Portfolios News Sentiment

Other key findings were as follows:

  • News sentiment worked everywhere, with positive performance in 13 out of 15 markets for Large/Mid-Cap and 14 out of 16 for Small-Cap
  • Hong Kong was the best performing exchange in the backtest for Large/Mid-Cap companies, with Annualized Returns of over 42% and an average portfolio size of 11 names
  • The study provided good evidence that news sentiment still brings value over longer trading horizons, with predictability of cumulative returns up to a month ahead
  • Nevertheless the strongest signals were for the 1-day horizon, giving good predictability for next-day returns. These were compared across the region by showing their Information Ratios in a heatmap below

APAC Heatmap News Sentiment IR Mid-Large Caps vs. Small Caps

Short vs. Long Sentiment Signals in APAC

Researchers also compared the performance of long vs short signals and found short signals outperformed long over shorter holding periods, however, over longer holding periods performance tended to favour longs.

“Negative news tends to be more extreme, and the more extreme the sentiment, the greater the market impact.” Says Peter Hafez, Chief Data Scientist at RavenPack.”

Sentiment Signal Decay in APAC

A further focus was how well signals performed over longer investment horizons, with researchers finding that on a cumulative returns basis signals they continued to be successful for up to a month ahead.

Indeed, for Australia & New Zealand the optimal holding period seemed to be about a week, whilst APAC as a whole exhibited IRs greater than 1 for Mid/Large-Cap and 2.5 for Small-Cap for holding periods of up to a month.

One of the conclusions of the study was that decay was slower for the Small-Cap than for the Large/Mid-Cap, because of the former’s lower liquidity and slower incorporation of news information.

Another clear observation was that negative signals, while stronger, also decayed faster. This phenomenon was so prevalent that in many cases, and particularly for Small-Caps, a crossover was observed when moving towards longer trading horizons, as positive signals started to outperform.

News Event and Sentiment Data on APAC Stocks

Easily implement the results of this White Paper using the RavenPack Analytics Platform, which includes a comprehensive and high-quality database of news event and sentiment data on APAC stocks, available via dashboards or our web APIs. Request a trial today.




We will process your personal data with the purpose of managing your personal account on RavenPack and offering our services. You can exercise your rights of access, rectification, erasure, restriction of processing, data portability and objection by emailing us at privacy@ravenpack.com. For more information, you can check out our Privacy Policy.