Innovation as quiet evolution
If you have a 20-year-old car, and over time, you’ve replaced several parts – the windows, air conditioning, steering wheel – is it still the same car? Wouter Tilgenkamp, Portfolio Manager, would argue it is. Over the last 20 years, Robeco’s approach has improved: factor definitions have been enhanced and new alpha signals introduced, while portfolio construction algorithms have been revamped – but the soul remains.
In some ways, innovation originates in the freedom to pursue goals with the sense things can be done better. Having a mix of stable, experienced team members and fresh perspectives from new hires or university graduates anchors that perspective.
New ideas are also explored in a controlled environment: Robeco’s incubator, where the team tests concepts with Robeco’s own money. Only when an idea has been analyzed, examined and proved mature is it implemented broadly.
Why communication matters
Communication and partnership are also key when it comes to innovation. It’s vital for asset managers to clearly articulate how and why their strategies are evolving, so that clients can follow that reasoning. Innovative elements need to be repeatable, explainable, and address genuine challenges.
Indeed, while clients understand traditional factors, the world has changed dramatically in the last decade. Staying ahead increasingly requires the integration of AI and related technologies into core operations. The addition of new signals is best seen as evolution, not revolution.
Descubra el valor de la inversión cuantitativa
Suscríbase para conocer las últimas novedades y estrategias sobre inversión cuantitativa.
Signals to watch for: Green flags and red flags
What are green flags and red flags when it comes to innovation in quant? Jeroen Hagens, Client Portfolio Manager, shares that a key green flag is the ease and competence with which quant teams handle new research techniques.
For example, existing strategies can be augmented with next-gen quant signals, including variables that combine alpha generation and sustainability elements, while machine learning approaches are being explored for stock selection. These techniques are good at identifying alpha opportunities while also managing risk.
Alternative datasets, says Hagens, also provide researchers and portfolio managers with a pool from which to fish unique insights. One recent variable Robeco derived uses NLP to translate earnings call audio into sentiment analysis. Another is a job momentum signal, which gauges a company’s health by tracking job vacancies and layoffs.
Red flags, on the other hand, include incorrect or unreliable data and overuse. Research teams are offered new datasets by vendors constantly, and each must be scrutinized to ensure the data isn’t from questionable sources or altered after the fact. Even reliable machine learning methods can be overused. In the end, it’s about getting the delicate balance right between new techniques and established foundations.
“We’ve seen exponential growth in data and computational power, and it’s hard to predict where we’ll be in five years,” says Hagens. “It’s exciting but also a bit daunting because the future is so open.”
Conclusion
Innovation doesn’t always look like a breakthrough – it often looks like refinement. In industries that prize precision and reliability, real progress is steady, explainable, and grounded in both data and discipline. Staying ahead doesn’t mean chasing every trend, but building the capacity to adapt – and to know when not to.
