As pointed to in our recent coverage, machine readable news is expanding its asset scope and geographical delivery. A good time, then, for IntelligentTradingTechnology.com to get an inside view on the topic from Rich Brown, head of Elektron Analytics at Thomson Reuters.
Q: What’s your assessment on the current state of play re. trading firms using machine readable news services as part of their trading strategies?
A: The use of machine readable news in trading strategies has been growing dramatically over the last few years, but its widespread adoption is still in its early stages. Increasingly, firms are incorporating it into their processes in a defensive posture, utilising it as a circuit breaker or as a way to inform of future volatility to better time trades. There is a lot of untapped potential in this space and as more and more firms begin to incorporate machine readable news services into their strategies, it will become commonplace and a must-have set of signals for even basic strategies.
Q: And the same question but about trading firms using social media sources?
A: The use of social media in trading strategies is in its infancy. While there may be many examples of breaking news first appearing in social media, the challenge lies in the amount of noise present in social media. The vast majority of users are not talking about stocks and for those that are it still takes quite a bit of work to determine whether or not those comments are tradable. When analysed properly, this noisy signal can be cleaned up and put to work against certain use cases.
Q: How much is sentiment analysis playing into trading decisions?
A: Sentiment is one of many useful ways to make trading on news and social media more viable. Combining this with other text analysis techniques like relevance, novelty, topic, intensity/buzz, and co-occurrence can significantly increase the signal to noise ratio and allow traders and investors to focus in on the specific market moving events that are relevant for their portfolios. It enables them to get more clear signals for future returns, price volatility, and trading volume allowing them to adapt their trading to the most current (and likely future) market conditions.
Q: Thomson Reuters have introduced a number of new services in recent months related to machine readable news, social media and sentiment. Can you briefly describe what each does, and how a trading firm might use them?
A: For event-driven trading, we recently launched an exclusive ultra low-latency release of the Institute for Supply Management Report On Business, providing a timely and accurate gauge of current business activity across the U.S. and global economies which is considered by many economists to be the most reliable near-term economic barometer available. These monthly indicators provide more frequent trading opportunities and more timely information compared to quarterly economic data.
In addition to the premium financial news from Reuters and select market moving sources, our News Analytics system now incorporates internet news and social media content, enabling users to incorporate information from some 54,000 Internet news sites and 4.5 million social media sites. It also helps you determine where you are in the information cycle. For example, is your novel trading idea already well known? Did this latest piece of information show up on social media first and how widespread is it? Is there more “information momentum” behind it, enough to ride the price momentum or exploit the reversal?
We have also launched Thomson Reuters MarketPsych Indices which measures psychological states of various countries and asset classes. These indices produce just over 18,000 data points every minute measuring things like fear, optimism, anxiety, sentiment, and others across agricultural commodities, materials and energy, currencies, equity indices, and countries. Increased levels of fear may lead to more safe-haven investing and measuring the buzz on regime change or levels of optimism in certain countries combined with currency chatter may give a global trader a number of ideas to exploit this information advantage. It also provides the psychological backdrop within which many of her/his other factors operate.
Q: With reference to the services you now provide, would trading firms generally use one of them, more than one, all of them? Are they meant to be complementary?
A: The services are meant to be complementary and allow trading firms to choose one or all of the services, depending on their strategy. A trading firm can use our machine readable news feed as a circuit breaker, but add our news analytics to refine which of the news items actually cause the algorithm to stop, or bias their market-making quotes in a particular direction, or even turn the algorithm back on after the circuit breaker when it is determined that the news item is not relevant or material to the strategy. As a trading factor, News Analytics also performs differently and can be adjusted by the information revealed through the MarketPsych Indices.
Q: What’s the near/mid-term future hold for news/social/sentiment driven trading? Where is the focus heading, in terms of functionality, delivery, markets covered?
A: In short, more data, more metrics, more markets, more languages, more offensive exploitation for leaders, more defensive adoption for laggards, more signal, and less noise.
At length, there is a significant amount of research showcasing how the data can be used for different strategies. As this industry evolves, one will naturally see a number of ways to continue enhancing the data sets, but the biggest value for most firms will be in how it works for their specific strategies or processes.
We see a number of ways for humans to be able to exploit the value of these techniques. Gone are the days where machine readable news is isolated to the high frequency, black box play. Stock screening tools with top 10 lists, pairs trading, idea generation, risk management and hedging, automated/customised alerts, support for asset allocation decisions, data visualisations and exploratory data analysis using the conclusions from the quantitative research will provide the human with “intelligence amplification” enabling them to take in and make sense of a wider variety of information in the market, from trusted sources like Reuters news to the citizen journalists of social media and beyond.