With robo-advisors and improved regulation, machine learning could make financial systems friendlier and more rewarding.
- Machine learning improves personal-finance services. In some cases, robo-advisors already perform better than their human counterparts.
- With all European banks forced to open their data gateways to external service providers by 2019, many new financial services based on artificial intelligence could emerge.
Automation has radically transformed the financial sector – and we’re not just talking about cash machines. From approving loans to managing assets to detecting fraud, human involvement in financial services is declining. Swiss bank UBS, for example, could shed almost 30,000 workers in the coming years because of these technological advances, says CEO Sergio Ermotti.
The use of automation and complex algorithms is most widespread in high-frequency trading. The majority of deals in equities and futures are now executed automatically by machines rather than on a trading floor – and at lightning speeds. “Thanks to algorithmic trading, you can make a profit on a very short timescale,” explains Mathieu Rosenbaum, chair of Analytics and Models for Regulation at École Polytechnique’s Centre for Applied Mathematics.
This evolution, however, has its downsides. “The market price of assets and their fundamental value can become disconnected,” says Rosenbaum, creating opportunities for market-wide bubbles, as happened in tech markets in 2001 and housing markets in 2007. Machine learning could provide a corrective. “New tools allow us to learn about the relationship between inputs and outputs in abstract terms, and then predict how outputs may differ as inputs change,” says Damir Filipović, chair of the Swiss Finance Institute at the École Polytechnique Fédérale de Lausanne. At the macro level, these algorithms may be used to improve financial regulations.
Recommendations and convenience
How can machine learning make financial systems friendlier? For the customer, machine learning is improving bespoke personal-finance services. “People find it difficult to plan their financial future because there are so many unknown variables,” says Filipović. “On the other hand, they are part of a population, and this is where data science can help. Everyone has characteristics or attributes that help classify their situation and infer recommendations relating to their likely future trajectory.”
MoneyFarm is one example of such a “robo-adviser” using machine learning to make financial systems friendlier. Founded in Italy in 2011, the company creates investment solutions designed to grow wealth inexpensively and with maximum transparency. Measured by assets under management, MoneyFarm is the third largest player in Europe, behind UK-based Nutmeg and Germany’s Scalable Capital.
Machine learning making financial systems friendlier
MoneyFarm’s platform deploys a computer agent that aims to deliver returns based on risk profile. These agents can perform better than their human counterparts: their return was 18.2% in 2016, vs. 13.1% for the average active managed fund, according to the company. Without brokers’ salaries to pay, MoneyFarm can offer significantly lower fees. And like Amazon and Netflix, financial personal assistants will soon understand a costumer’s needs better than they themselves do. As in those industries, convenience may prove to be a decisive advantage.
Swedish start-up Minna Technologies lets people manage, switch or cancel existing utility and software subscriptions through a single app. By training its algorithms on population-level datasets, the app represents an increasingly useful personalised financial-management tool. The company says it has saved customers €15 million since its launch in 2016.
More than 200,000 Swedes use the service, and the company is now set to partner with Danske Bank, Denmark’s largest. A few years ago, banks seemed locked in battle with fintech start-ups; now collaboration typifies the landscape.
“In five or 10 years, much of the infrastructure you see in banks will be completely different,” predicts Kourosh Marjani Rasmussen, associate professor of operations research and financial engineering at the Technical University of Denmark. “The smart banks offer access to customers and data, and in return they leverage knowledge from the start-ups to ensure they’re part of the transition from the old infrastructure to the new.”
Pragmatic companies like Denmark’s Spiir are demonstrating how partnering with banks opens opportunities to innovate in the face of regulation. Since 2011, Spiir has helped Danes manage their budgets, monitor their spending and reduce fixed expenses from their phones. Recently, the company secured €3.25 million in investment from Danske Bank, partly thanks to its new platform, The Nordic API Gateway, which gathers data from more than 250 banks. The platform gives Nordic banks and third-party providers a head start on the EU’s revised Payment Service Directive (PSD2), which requires all banks to open their data gateways by 2019.
With PSD2 soon taking full effect and robust levels of investment in fintech – 612 venture-capital investment deals, worth €2.7 billion, were struck last year – Europeans could soon see a host of financial services that use artificial intelligence to place individuals back at the sector’s heart.
First published by Technologist magazine