Alternative data sources for financial services has been getting a lot of attention recently. Like any trending topic, its uses and benefits can be hyped up a bit. Much of the attention has been around creating trading strategies with an edge (generating alpha). But it also applies to risk management and enhanced regulatory oversight. To better understand the utility of alternative data, this post presents the topic within a larger context that will help to put alternative data sources to use in the most efficient and practical manner possible.
Alternative Data is a Relative Term
Alternative data sources could be any type of data, depending on your point of view. What’s unique or new to one will be common and well understood by another. Below is a general list of alternative data types as it relates to financial services. It is not a complete list because new types of data will be discovered while existing ones become mainstream.
- Satellite imagery;
- Geolocation data/location intelligence;
- Logistics data;
- Social media and news feeds;
- IoT captured data;
- Weather trends;
- Data enriched by machine learning or artificial intelligence.
Alternative data is often associated with external and exotic data sources. But internal data can also be included in this category. For example, business performance information about companies is used for evaluating counterparty risk as a normal course of business. It is traditional and well understood data for the risk team. However, to a trading desk within the same organization, that data could determine whether to go short or long on a manufacturer based on the health of the companies throughout its supply chain. In the regulatory space, leveraging electronic communications such as email, chat, and voice to enhance trade surveillance is also using internal data (e.g., an email server) but in an alternative way. An open mind can find new sources of alternative data even if it already belongs to you.
Having provided examples above, let’s take a quick look at how alternative data sources could be applied.
|Application||Example (alternative data source in bold)|
|Capture Alpha: Use unique data sets that by themselves, or in conjunction with traditional market data, influence trading strategies with additional insights in order to gain a competitive advantage.||The geolocation of cargo ships and weather analysis indicate late or non-delivery of goods that could lead to a localized spike in prices. This is insight provided prior to reporting by local markets.|
|Enhance Regulatory Oversight: Leverage non-market data to build a comprehensive picture of events in order to optimize the identification, investigation, and remediation of compliance risks.||Data enriched by machine learning routines identify suspect trade flow patterns at a very granular level. Then, specific transactions are linked to newsletters and messages that have been analyzed for misleading statements. This provides a deeper understanding of both the event and the intent of a price manipulation scheme (“pump and dump”).|
In summary, alternative data could be anything, depending on who is using it, and applied in any way.
New and Useful for a Specific Purpose
Two points that are critical to making use of alternative data are as follows:
- Alternative data are data sources that are both new and useful for a specific purpose.
- The purpose of alternative data is to provide additional and actionable insight to a strategy before it becomes well understood and generally used.
“Now is the time to utilize these new data sources to gain an edge. An edge can only be found via information that only one or very few firms have. Therefore, once an alternative data set becomes ubiquitous, the edge it provides will start to slip away.
–Kevin McPartland – Greenwich Associates
The competitive advantage diminishes and alpha decays when a novel method becomes generally understood and adopted. Likewise, in the regulatory space, surveillance becomes less effective as bad actors devise new ways to cover their tracks based on the knowledge of how others got caught. Alternative data has a shelf life.
The shelf life of alternative data is extended by the sound execution of a plan or design. If competitors have access to the same alternative data sets, but their implementation is poor, an edge is still retained.
Driven by Subject Matter Experts
Alternative data is challenging because: (1) it requires creativity and discovery; (2) it is unprecedented, so there may not be a template to work from; and (3) time is limited. How does one go from an inkling of an idea to generating real value while the window of opportunity is still open? The answer is to enable subject matter experts to drive the solution.
“RegTech: Leverage Alternative Data for Compliance”
A common theme across the financial services regulatory and compliance landscape is to provide a robust and defensible process to the regulators. To do that, you need large volumes and many types of data. To that second point, see this webinar about leveraging all types of data to enhance your regulatory programs.
“RegTech: Leveraging Alternative Data for Compliance” is a panel discussion that covered lots of ground with some great insight from RBC and Nordea. They give very practical advice about how to kickstart your use of alternative data.
- Start with low hanging fruit
- Make it a collaborative effort
- Leverage public data sets
- Expand gradually from there
The purpose of alternative data is to provide additional and actionable insight beyond the expected / usual / commonplace.
The subject matter experts can put alternative data to use because they understand how to derive meaning (and thus value) from new types of data. The efficiency and efficacy of their work will also increase through collaboration. For example, a subject matter expert (e.g., a portfolio manager) can benefit from the bigger picture knowledge of an analyst, quant, data scientist, and other colleagues. She can take her idea and test assumptions that either support or debunk it. If a positive outcome, then continue the exploration and derive an application. If a negative outcome, move on to something else.
An Iterative and Exploratory Process
The process to derive meaning from new and alternative data sources is iterative and exploratory in nature. Experts need to be able to expand on an idea by experimenting with any type of data at any time. Throughout the process they should be able to discard sources that don’t fit and retain ones that do.
Native visual analytics enables subject matter experts to drive solutions because it enables them to run analytics where the data resides. You need a technology that runs natively in a big data platform, whether that’s Apache Hadoop clusters, the cloud, or other modern data platforms. Native visual analytics unifies data discovery, business intelligence, and real-time visualization in a single, integrated platform. The subject matter expert is able to directly leverage big data capabilities through an intuitive and self-service interface.
The ability to run analytics where the data resides is what differentiates native visual analytics from traditional business intelligence and visualization tools. In the native case, data stays in place, even if you are working with multiple data sources. The analytics and visualization engines are distributed at each source, simultaneously. In the traditional BI case, data sources are moved to a separate place before the analytics are run. Thus, results are dependent on getting that move right in the first place. In this case, subject matter experts are not driving the solution; they are waiting to make decisions.
Subject matter experts are best equipped to derive meaning from new and alternative data sources. But they need to do this before the data becomes well understood and generally used by the competition. The challenges of alternative data are reiterated below, along with the capabilities needed to address them.
- It could be any type of data, depending on the point of view: The subject matter experts must have the ability to explore and evaluate any type of data (structured, unstructured, real-time, or historical). Creativity is key to spotting and implementing new opportunities.
- It is unprecedented so there may not be a template to work from: Data sources must be accessed as needed to see how they align (or don’t align) regardless of the original purpose of that data source. A user needs the ability to experiment, unplugging sources that don’t fit while retaining ones that do.
- Time is limited: Real-time analysis juxtaposed with historical evaluations greatly enhances the ability to monitor, measure, and report changes to models. The ability to collaborate within a team leads to well-informed iterative improvements. This extends the shelf life of all related alternative data sets.
Interested in learning more about how you can put your alternative data to use? Check out our alternative data resource center.