Access to data is crucial but one must not entirely rely on it, SRP Europe

Having access to data is crucial, but one must not entirely rely on it, according to panelists during the How Technology Will Change the World of Finance discussion at the 15th Annual Europe Structured Products & Derivatives conference at the Etc.venues, County Hall, in London on February 8.

One of the specific challenges investment companies face is the process before they can distinguish valuable data from useless data, according to Daryl Smith, director of research, Neudata. "Around 70% to 90% of this process involves cleaning, sorting and processing the data until you reach the state when you can back-test it," Smith said noting that another common issue is the lack of history that would allow companies to back-test the data. "Some of our clients like unstructured data that is tricky to process because they see it as entry barrier for their competitors," Smith added.

Unstructured data is the scene of revolution right now, said Johan Groothaert (pictured), CEO and co-founder of non-bank lender Fiduciam. “With machine learning taking over data handling, it is now possible to process greater amounts of data than before and make sense out of them,” Groothaert said.

According to Jamie Smith, head of product at Smart Pension, the UK-based platform for automatic pension enrollment, the benefit of big data is “about finding something that you can use as an advantage and that must not necessarily be a breakthrough”. “Gathering large amounts of data that is also accessible to others, then investing time to work with it only to reach the same conclusions as others is type of efficiency that naturally comes about, but brings no true value,” Smith said.

Huge amounts of data are not a must if you have a clear objective in mind and try to test a hypothesis by understanding the data rather than just go through big volumes of it, according to Smith. “Furthermore, the data itself cannot be helpful if one does not understand the underlying [human] behavior and why it is of interest.”

Just like in the 90s when the earnings-to-price and momentum strategies were found to outperform the market, today’s US credit data shows really good results, said Groothaert. “Humans are still much smarter than computers and can manage their own credit score to appear better than it really is,” he said. “One must understand why particular strategies work when they work because they may not always do.”

“What structured products and big data have in common is that their definitions are both extremely broad-based,” said Daryl Smith. “The fact is that the data providers out there today didn’t exist one or two years ago”, Smith added.

Robo advisers and technological capability are already here, according to Jamie Smith. “We only need to get pass the threshold where customers are ready to talk to a robot instead of a human financial adviser”, he explained, “because until now, the trust element has been linked to human interaction, especially in the context of emotional financial decisions, such as investing one’s life savings”. “It is only public acceptance of computers’ ability to deliver the same information as human advisers that will lead to robo advisers becoming mainstream.”

In terms of regulation, the way of dealing with data and understanding the clients and their objectives is of great importance and heavy regulation within the “rapidly developing environment” of investment technology can do more harm than good, said Groothaert.

According to Ryan Shea, head of research at Amareos, it is not about the story of humans versus machines because in the end, technology-based investment vehicles, such as artificial intelligence (AI) funds, cannot exist without human supervision. “It is the human-machine interaction that is very powerful.”

  • Published February 2018
  • Author Darina Bacheva