Sophisticated investment managers learned long ago that its more and more difficult to find actionable info before the market can react to it from traditional sources of financial data – such as SEC filings, earnings calls, and quarterly statements. Gaining that edge increasingly requires harnessing alternative data, which can run the gamut from the fairly basic to, well, literally anything.
In recent years, this motley category has grown to include seemingly every type of information imaginable, including consumer behavior, credit card transactions, product reviews, web traffic, app usage statistics, weather information, social media posts, data from objects enabled with connectivity (IoT sensors), corporate travel, and even satellite imagery, to name just a few examples amid thousands.
The challenges in gathering and analyzing such data are also legion, of course. It comes in countless forms from disparate sources, is typically dirty and complex, may be inaccurate, and may have even been obtained illegally – all of which creates perils for investment managers attempting to utilize it. Overcoming these hurdles is expensive and time-consuming, and requires deep expertise in data science and significant technology resources. But the competitive benefits in finding alpha first with this info can be worth it.
“Even if the data is not perfect and there’s a lot of noise, the fact that you can capture information faster and make use of it allows you to position your portfolio before investors relying on traditional financial and economic data,” says Laurent Laloux, Chief Product Officer of Capital Fund Management (CFM), a global quantitative and systematic asset management firm founded in 1991 and a leading firm in using alternative data.
Getting a jump on the macro and micro
To gain advantages in macro strategy, the right alternative data can give an investor an early reading of country or continent-wide economic developments, or issues affecting big asset classes, says Laloux, by analyzing info regarding supply chain changes, early shifts in consumer behavior regarding commodities, and so forth.
At the other extreme, when alternative data is scrutinized and analyzed carefully, it can give an indication of a company’s status or intentions before they make such information more widely available. “If you can measure consumption and supply-chain factors impacting a specific corporation, you can assess its economic situation well before the earnings call which is dated,” says Laloux. “By the time a corporation publishes earnings, it’s really two months after the battle has been fought. Only the investor who had that information ahead of time could really put it to use.”
Effectively harnessing the growing volumes of alternative data to isolate new signals and create investment opportunities is inherent to CFM’s mission, Laloux adds. The firm utilizes 6+ petabytes of data that’s constantly refreshed to manage its portfolios, executing 130,000 trades daily with $11 billion in AUM.
“We invest across thousands of assets, and we leverage the statistics of doing that and collecting as many data sets as possible,” adds Laloux. “By aggregating all this information, we get advanced information that provides added value to all the sectors of our portfolio. Many of the clients that invest with us find it challenging to provide the time, staff expertise or infrastructure to process complex data sets at scale, and they look to us, as an experienced quant manager, to create that value for them. “
The need: Real-time, holistic, 360-degree view
Speed and scope are the key logistical aims of an alternative data solution, Laloux emphasizes. To consistently create value, you need a real-time intelligence from a comprehensive spectrum data that doesn’t lack any vital areas. “Your alternative data strategy needs to give you 360-degree understanding of the real-time economy, so you can know quickly as possible what is currently happening within a specific country, segment of consumers, corporation, or any other kind of economic agent,” says Laloux.
To achieve this, he adds, automation is essential. “It’s critical for your alternative data solution to be as automated as possible,” says Laloux. “In being systematic and quantitative, all of our monitoring is automated so that we can extract the key performance indicator (KPI) from the data quickly and incorporate it into our strategies for better outcomes. If this process is not well-oiled and is inefficient, you lose those opportunities.” Advanced automation is especially critical to scale up and manage bigger and more complex data sets, he notes.
Choosing a data provider: 4 vital points
For asset managers with sophisticated data science capabilities that choose to procure and analyze alternative data in-house, most journeys start with selecting alternative data providers who can supply the firm with data sets.
“Sourcing and qualifying the data providers at the early stages is really critical, and this requires a lot of manpower and know-how because it’s not an easy task,” says Laloux. “There are big providers, small providers, niche providers, firms that are familiar with institutional investors and asset managers, and others which are totally remote from the financial world and don’t even understand the type of questions we ask. You need to have a due-diligence list you can ask the provider the correct questions and, as soon as there is a red flag, eliminate and focus your attention on others.”
While due-diligence checks are extensive, Laloux says the biggest trouble often comes from a few common areas. For investment firms that want to navigate these waters, he recommends that they give special attention on their due-diligence questionnaire to these issues.
- Legality. “You need to ask, ‘what is the legality of the data they are providing? Can you safely trust the provider and consume this data?’ Given the different regulatory landscapes, this issue is increasingly critical.”
- Point-in-time rigidity. The data provider must be extremely clean about maintaining data exactly as it was at the time it was first available, and not polluting it by adding information to it at a later point. “Data providers may try to explain this away by saying, ‘I’m just cleaning the data because I’ve noted this mistake.’ But if this mistake wasn’t known to the market at the original time, they’re massaging the data – and that’s a critical red flag. There must no information from the future put back into the data, otherwise it puts you at risk for future look-ahead bias.” Fortunately, there are several systematic tests that can check a data set for point-in-time cleanliness, so you can verify this essential characteristic.
- Compatibility with your own data. Your due-diligence questionnaire should probe for details on how the provider distributes and references their data, being specific about the format, shape, and naming choices. “If a data provider gives me ‘International Business Machines’ as one big string in their data set, but my internal data referential is the ticker, ‘IBM,’ how do I connect these two entities so that I can do my job in the end?” asks Laloux. “That seems trivial, but given the complexity and the diversity of providers, that’s a real hurdle to having an efficient process.”
- Cloud technology. “We are seeing a big wave of advancement in cloud capabilities across technologies,” says Laloux. Firms should choose alternative data providers that are making full use cloud technology to distribute their data. If your firm already uses a large cloud service, that company may be able to help you scout this territory. “They typically give you a catalog of the data marketplace, so it helps you know which firms can supply data that will connect with the technology you already have.”
The ultimate test: Is the alternative data adding value?
After choosing an alternative data provider, investment firms are often allowed a short period – from a few weeks to three months – to test the data set before deciding to buy it. To do this effectively, firms need to have an established testing procedure so they can evaluate a data sets quality and value to their investment decisions in a brief period, Laloux notes. Most asset managers don’t have this testing protocol at the ready, and don’t realize they need it until they’re into the trial period, he adds. This creates an additional pain point in utilizing alternative data for many firms, and it increases the chances that they’ll make an uninformed decision in purchasing access to the data set.
“A typical issue is, what you see in the test is not what the data provider gives you in the real-life setup,” Laloux warns. “There might be discrepancies, so you need to be wary of that and know exactly what you’ll see when the data set is coming through to your production setup.”
As mentioned earlier, speed and scope are two of the more important characteristics you need to assess the quality of an alternative data solution, and you should quickly ascertain if the technology is allowing your investment team to quickly get a real-time, holistic, 360-degree view of the entities they need to focus on (from countries to single companies) in order to act before the market and achieve the goals of the portfolio.
Indeed, as only results truly matter, Laloux stresses that your testing process must allow you to confidently answer the ultimate question: Are you extracting value out of the data?
“Does the alt data contribute positively to the performance of your fund? That’s the entire point. At CFM, we ensure that the team that’s processing and analyzing the alternative data has ownership of the models, and on a regular basis they can ask, ‘how is this data set contributing to CFM funds and to investors’ alpha in the end?’”
CFM takes a long-term view
Unfortunately, many alternative data sets that deliver useable information at the outset don’t maintain their value after a period of months or years, as its signals become better known in the market at large or simply erode in quality, causing its alpha to decay. While some investment firms burn through multiple alt data sets after using them more superficially for brief terms, CFM looks for much deeper engagements with higher-quality alternative data sets.
“We use data sets that we believe will add value to our portfolios over the long-term, not over a horizon or just a few months,” says Laloux. “We use data sets that we can fully understand, integrate, and find complex relationships within – not only as standalone data entities but also in the perspective of cross-referencing them with other data sets globally, so that we can generate an even wider understanding and leverage this information across all the teams internally and across all our strategies.”
This also allows CFM to build relationships with their alt data providers, so they can provide valuable feedback and understand how the provider is evolving. As each renewal date approaches for each provider, Laloux explains, CFM teams will confer on the quality and effectiveness of each data set, and collectively decide if it still provides the necessary value to the portfolios to justify its cost – or needs to be replaced with a superior data set that can deliver better results
“When you buy hundreds of alternative data sets and build hundreds of models with them, as CFM does, you need to make sure that all the data remains fresh and useful, and continues to add value to the portfolios,” says Laloux. “While we seek long-term relationship with alternative data providers, we won’t hesitate to change data sets when they no longer generate alpha for our investment teams.”
“Baked into our process”
Importantly, Laloux notes that all of the vetting principles for data providers, testing goals for alt data solutions, and other recommendations he offers to asset managers seeking to harness alternative data to gain competitive advantages have long been an inherent part of the investment process at CFM.
“The innovative strategies that we generate from these alternative data sets are already baked into our research and decision-making process at CFM, and are currently live and successfully working for us in our products,” he says. “And this is a key differentiator for CFM; not all other quant firms and investment managers have the same abilities with alt data, and not everyone has three decades of experience in turning huge quantities of data into successful outcomes for investors.”
As for the future at CFM? “It goes back to our culture. Since 1991, we’ve been growing with technology, we’ve been growing with the availability of data in all its forms, market, financial and alternative, and we will continue to do so,” Laloux says. “That’s really our DNA at CFM.”
Learn more about harnessing alternative data to find alpha.
Disclaimer
Any description or information involving investment process or allocations is provided for illustration purposes only. There can be no assurance that these statements are or will prove to be accurate or complete in any way. This article does not constitute an offer or solicitation to subscribe for any security or interest.