Researching, designing and implementing top-notch quantitative investment strategies requires good data sources. In order to ensure that all the necessary data is available to all quant research and investment teams, Robeco has dedicated considerable resources to upgrading its data platform over the past couple of years. We talked to Jacob Buitelaar and Richard Groenewegen about these improvements.
Jacob Buitelaar: “Our aim is to ensure that Robeco’s quantitative teams have top-notch data infrastructure at their disposal. The key idea for that is to bring the research and the production data together in one place and make it available directly and flexibly to all relevant stakeholders, including researchers and portfolio managers.”
“Doing so, we can significantly shorten time-to-market for our clients. Once we’ve researched investment ideas, we can put them into production really quickly. Moreover, making data available centrally to everyone improves the data quality and the efficiency and collaboration.”
Richard Groenewegen: “It starts with historical data concerning financial markets, like asset prices or transaction volumes. But it can also be sustainability-related information, or proprietary data regarding our funds’ positions and holdings, as well as transaction costs, for example. In all cases, we check the link between each piece of data and the related financial instrument. So, it’s a lot of details.”
Our teams may need to process a lot of data and we want to be able to scale our computations
R.G.: “Definitely. In fact, that’s the other part of what we do: we make sure that we also have the necessary processing power and the tools to analyze the data. Our teams may need to process a lot of data and we want to be able to scale our computations. That’s why we’re moving things into the cloud, where we get access to some of the most advanced tools, such as Apache Spark.”
“The cloud allows you to effectively clone or duplicate your environment, which makes it easier to have research and production all effectively on the same platform without risking any integrity issues. And because all the different teams work with exactly the same tools and the same data sets, it makes the transition from research to production much easier.”
J.B.: “Altogether, this means much more than having interconnected teams: the whole process works in exactly the same way from beginning to end. This enables both scalability and flexibility, something increasingly important as the data sets we use as input for our quantitative strategies are becoming larger, more diversified and often less structured. Just think about SEC filings or newsfeeds, for example.”
R.G.: “Yes. For instance, take the cloud implementation of Robeco’s quant allocation strategies. This project had two main aspects. The first one was to have a front-to-back automation of the quant model, starting with data collection, all the necessary checks and putting the monitoring of the process in place. The second aspect was to work with new technology and to bring IT development much closer to the front office.”
“So, it was really about having a team with one foot in the front office and one foot in IT. And now, instead of running scripts and babysitting the whole process, we simply run everything automatically, with data-anomaly checking and web dashboards in place to ensure confidence in the outcome.”
“Another project we worked on during the second half of 2019 was a bottom-up reimplementation of the quantitative selection database. This database gathers stock-specific information, like historical prices, transaction volumes, company earnings estimates, accounting information, news sentiment information, ESG information, and so on and so forth. We have put a lot of focus on making it flexible, so that we can quickly add new data sources. ”
J.B.: “This is a very important aspect: we want to be very flexible and able to add new data sources really quickly. For example, if a researcher needs a new data source to be included, we want to be able to have it in the next day. And we want it to be available not just locally but for the entire research department, which requires huge processing power and cutting-edge technology.”
“In particular, we use the Databricks platform and its analytics engine Apache Sparks. Doing so, we can have hundreds of computers available within five minutes working on the same problem. That’s several terabytes of main memory that we can use flexibly. We also work very closely with Databricks to ensure that features and performance of the platform that are specific to our use case get the right attention on their development agenda.”
R.G.: “We may slightly shift our focus towards the computation side of things to enable faster processing of large datasets as well as machine learning. We’ll also likely add more data sources, including further sustainability-related information or insights from our fundamental analysts, for example. Also, we will be looking to extend the quantitative selection database with credit data and even more detailed information.”
J.B.: “Not at all. In fact, this is one of the key benefits of our new platform: instead of having separate infrastructures for each capability, we gather everything in one place. Because our investment strategies combine cross-asset signals, it is really useful to have them all together. This way we can link instruments and use signals from one asset class or the other.”
R.G.: “We are also helping to create a tool to automate the customization of Robeco’s Core Quant Equity strategies. This would enable clients to configure a personalized portfolio reflecting their risk, return and sustainability preferences in real time. It would allow them to understand better tradeoffs between various options and to see the effect of changing some existing portfolio settings.”
R.G.: “Exactly. We observe a trend that once investors have chosen their index, they are looking for the best way to enhance that index as an alternative for a pure replication, or passive, approach while staying close to that index. With this tool, clients can see directly the impact of their choices on the portfolio characteristics.”
BY CLICKING ON “I AGREE”, I DECLARE I AM A WHOLESALE CLIENT AS DEFINED IN THE CORPORATIONS ACT 2001.
What is a Wholesale Client?
A person or entity is a “wholesale client” if they satisfy the requirements of section 761G of the Corporations Act.
This commonly includes a person or entity: