When Intelligent Data Becomes a Game-Changer
By Dr. Jan-Carl Plagge, Head of Applied Research, STOXX Ltd.
The investment industry has always relied on information patterns and trends to boost returns, with items such as regulatory filings and analyst reports long considered to be the traditional and established sources of data.
However, in the last few years, the volume of data available has grown in an unprecedented manner, and so have the channels producing it. It was estimated in 2013 that 90% of data in the world then had been created in the previous two years alone.1 Much of it is derived from the Internet of Things (IoT) and has led to the rise in use of what is termed alternative data.
So what is alternative data?
Alternative data is not a new concept. A myriad of unofficial and non-conventional sources of information have been used for years to determine whether an investment is a buy or sell. But investors who use the latest data streams are raising the stakes by looking in hitherto unexplored places, and widespread intelligent use of new sources of information will have far-reaching effects.
A recent report by J.P. Morgan2 classifies alternative data into three segments depending on how it is generated:
- Individuals – includes social media posts and search trends.
- Businesses processes – varying from company exhaust statistics to credit card transactions.
- Sensors – information such as ship locations and satellite images.
Some of these are already very popular. According to Greenwich Associates and Arcadia Data,3 over 60% of asset managers are currently using social media and social-driven news feeds as part of their investment process.
These sources of information were not originally designed for investing purposes. However, financial firms are finding ways of turning the data into sophisticated investment tools. For example, to gauge business sentiment on a global scale, the BlackRock Scientific Active Equity Team adds company conference calls to their existing data sources. Machine-reading algorithms ‘listen’ to the calls to track the occurrence of specific terms mentions, whether positive or negative, when discussing the outlooks for countries, sectors or markets.4
Machines and algorithms needed
While the abundance of high-quality data can be a pool of valuable investment information, sophisticated analytical processes are needed to organize and analyze it, and determine its applicability and utility. Here is where artificial intelligence (AI) techniques and so-called machine learning play a key role. Both are the key to construct algorithms that identify data patterns and make predictions.
In brief, AI is technology that behaves and thinks like humans. Machine learning is a subset of AI, and it denotes an automated process of cognitive detection of patterns and past experiences to ‘learn’ outcomes. Machine learning is used, for example, in self-driving cars and in web search functions. Both machine learning and AI integrate information seamlessly from a variety of sources into different systems.
Faster computer speeds and large-scale cloud storage mean that big data applications that use AI software have the ability to quickly identify and measure risks from staggering volumes and velocity of information.
Putting the data to work in passive investments
As such, products where the underlying processes are driven by technology and big data can be customized into ‘intelligent’ investments that generate smart financial performance. The use of alternative data has been of particular significance in the realm of passive investing, where traditional market-weighted indices are evolving into new, data-driven and targeted products based on quantitative and qualitative analysis. These are designed with methodical precision and are generating countless new possibilities for investors.
The science of predictions and forecasts in investing will still be a challenge in terms of costs and resources. Market participants therefore will need to invest heavily in the required technology infrastructure and human skills. As a result, quantitative passive strategies have become increasingly popular as a more cost-efficient option.
STOXX, in its role as an investment intelligence hub, is collaborating with the best data providers in each field to provide the right tools. For example, the iSTOXX FactSet Thematic Indices are based on FactSet Revere data, the iSTOXX Europe Factor Indices was developed with Alpha Centauri and the STOXX Climate Impact and Climate Awareness Indices is based on CDP data.
By defining systematic and smart benchmarks, data-driven decisions can be incorporated into portfolio construction via indexing approaches. These simple yet powerful products enable investors to tackle the threats and opportunities presented by today’s interconnected global economy, while optimizing their investment objectives.
We will be reviewing AI and machine learning in forthcoming articles on PULSE ONLINE.
1 ‘Big Data, for better or worse: 90% of world's data generated over last two years,’ SINTEF, May 22, 2013.
2 ‘Big Data and AI Strategies,’ J.P. Morgan, May 2017.
3 ‘Putting Alternative Data to Use in Financial Services,’ Arcadia Data and Greenwich Associates, September 2017.
4 ‘Constant Change, Consistent Alpha,’ Blackrock Scientific Active Equity Team, October 2015.
You want to receive our PULSE ONLINE mailing that updates you on new articles? Please send us an email to email@example.com. You can share your feedback, comments or questions by sending an email to the same address.
Please fill the form below to attend :
When Intelligent Data Becomes a Game-Changer
PULSE Online Newsletters!