There is no shortage of new technologies that can improve retirement outcomes for pension fund members
- Blockchain and machine learning dominate the discussion
- These are complex technologies in an early stage of adoption
- Real time data analytics and common platforms are already operational and can save time and money
- The adoption of new technologies is critical to engagement
When it comes to technological innovation in pensions the two buzzwords are blockchain and artificial intelligence. Blockchain is potentially a revolutionary technology that could significantly reduce the costs associated with pension administration and custody. Artificial intelligence – or more specifically machine learning tools – also promises to optimise many areas of the industry. They could be used to improve communications with pension fund members or to deliver better investment returns.
Considerable investment is required to develop blockchain applications to their full potential. Such investment cannot be carried out by individual pension funds, or even by individual for-profit organisations. Large consortia are more likely to succeed. One prominent example is R3, a company backed by a consortium of over 200 global financial services firms. The company is developing an open-source distributed ledger platform as well as a commercial equivalent.
Machine learning is closer to becoming a mainstream tool in pensions, according to Adam Jones, chief technical officer at UK consultancy Redington. This is because companies such as Amazon, Google and IBM allow access to their machine learning tools, through strings of programming code called application programming interfaces (APIs). “It means you don’t have to be doing machine learning research in order to use machine learning. We are starting to use aspects of machine learning in our software, particularly the way we search for things, combine data together and draw conclusions. It becomes part of the toolkit that software development can use,” says Jones.
Machine learning thrives when applied to large databases, which are becoming increasingly available in the financial services industry. An important piece of EU regulation, the Payments Service Directive (PSD), is contributing to this trend by making payments data accessible. In its scheduled update, due in September, the directive will provide further opportunities for fintech companies to acquire and analyse masses of data on retail banking transactions.
This is where shrewd data scientists could use machine learning to gain insights on people’s financial needs, according to Don van der Steeg, strategy and innovation manager in the pensions, actuarial and insurance services team at PwC in the Netherlands. Van der Steeg says: “It is something that could transform the whole concept of financial planning. Data on transactions can be very telling in terms of your future financial well-being, of which retirement benefits are an important part.”
It is fair to say, however, that machine learning is still embryonic. But pension technology has come a long way, according to Raj Mody, global head of retirement and pensions consulting and leader for transformation, technology and investments at PwC. There are several examples of technologies that have been operational for years and while many pension funds might see them as innovative, or even futuristic, they are already delivering better pension outcomes.
Mody divides the world of pension fund technology into three categories: calculation and analytics; common platforms and; ecosystems.
Real-time analytics tools have been around for some time. Mody says: “It is odd to think that, in bygone times, if you wanted to update the analytics you used to inform your decision making, that often took several weeks, if not months. You often had to ask an adviser to work out those numbers and the process was long and complicated. It is now commonplace for decision makers to be able to access those analytics in real time when needed. That is an area that has rapidly transformed in the last five years.”
Tools such as Skyval, a web-based platform, offer real-time analytics as well as shared access to data. The platform was launched by PwC in 2013 for defined benefit (DB) pension funds and their advisers to solve an efficiency problem. Mody says: “Trustees, sponsors and their advisers are often repeating overlapping analysis and this is incredibly inefficient. It slows down decision making and introduces the possibility of error. Adviser-agnostic platforms like Skyval solve that problem by doing the number crunching once and for all.”
The concept of ecosystems goes a step further. Mody explains: “Suppose a pension fund is trying to do a buyout with an insurance company, and it needs to get a price quote. Wouldn’t it be better if all the insurance companies that want to bid for the process could confidentially access a common valuation platform, and each of them apply their commercial pricing assumptions to a benefit structure that has already been calculated?”
Skvyal offers that possibility, too. The process can save each insurance company calculating the structure of benefits for the scheme separately, which should lead to lower costs. It would also mean more companies are likely to bid. As a result, DB pension funds would potentially get a better deal for de-risking transactions, according to Mody.
Redington has also invested in technology, with the goal of improving client access to advice, research and data. It has developed software called ADA that allows pension funds and investment professionals to access data and run real time analytics on portfolio. The software has also modules for risk analysis, defined contribution (DC) pensions analytics and cost analysis. Furthermore, it helps investment teams to pool quantitative and qualitative research data.
The software was originally developed as an internal tool. It became clear there was an business opportunity in making the software available to clients, not just pension funds but wealth managers and other clients, too. Jones joined Redington in 2017 to develop this business line.
“So far, the pensions technology landscape has been dominated by quite technical but quite complicated software. ADA is focused on trying to democratise access, to make it easier for people regardless of their background to use technology,” says Jones. The software offers a variety of functions and other software developers can plug into it to build other tools. Jones also says enhancements are being planned, including functions that will help pension schemes with governance tasks.
Mody points out that the behavioural consequences of such innovations are already showing. “The fact that trustees can access analytics on a daily basis does not mean they should do it. That behavioural issue should be dealt with,” he says. This is a reminder that technology, while solving problems, can create new challenges. It is easy to see how the existence of tools that monitor pension fund performance at all times poses questions on governance.
But, generally speaking, there is more to gain than to lose from implementing functional, tested tools that help with mundane tasks. It would appear that many pension funds are already lagging behind on this. A 2018 PwC survey on pensions technology showed only 30% of employers were aware of pensions analytics platforms.
The survey highlighted a lack of awareness of many tools, from automated member communication tools to pension administration platforms. This is not to suggest the pensions world is at odds with innovation. Like many sectors it is probably just a question of slow adoption. But there is one kind of ‘everyday’ technology that funds are struggling with, according to van der Steeg.
That technology is social media. Van der Steeg explains: “Pensions have now become a relevant topic of discussion for people below 35 years of age. This is a major shift and pension funds simply do not have a good answer to that. They are professionalising many aspects of their communication and engagement strategies, but they are struggling to be relevant with that younger cohort, due to their lack of experience.”
Given that young people are users of social media, that technology is part of the answer. The goal for pension funds is not just to be relevant to young people but to facilitate financial education and engagement in order to deliver better member outcomes. Van der Steeg argues that pension funds can learn from other profit-driven industries how to use social media and other tools to engage members.
Yet some of the biggest innovations are taking place in member engagement, partly thanks to the availability of machine learning tools. Pension giant APG is sponsoring Kandoor, a platform that provides answers to financial planning questions. The platform is based on a combination of an algorithm based on natural language processing with the knowledge provided by more than 200 experts.
“Trustees, sponsors and their advisers are often repeating overlapping analysis and this is incredibly inefficient. It slows down decision making and introduces the possibility of error” - Raj Mody
The platform is free of charge and independent, since the experts are volunteers. Users can create an account on Kandoor and ask financial planning questions, which will be automatically answered by a chatbot. If the algorithm is unable to answer, it is forwarded to the experts and answered within hours. The algorithm then adds the question and answer to its database, improving its capacity to answer in the future.
For legal reasons, the answers stop short of providing actual advice. But this is an experiment to suggest that natural language processing, a type of machine-learning tool, can solve actual problems. So far, Kandoor has answered nearly 153,000 queries, according to its website.
It also shows under which conditions technological innovation works. Van der Steeg comments: “The interesting element is that this is not the application of a single technology. This is the combination of several technologies, a chatbot, a natural language processing system, with a community of coaches.”
In other words, technology of itself does not solve any problems. The combination of different technologies and a human element, with a focus on solving one issue, in this case the provision of free independent guidance on financial planning questions, is more likely to be successful.
That is why the implementation of new technologies will inevitably consist of a series of discrete interventions, rather than a big systems upgrade, according to Jones.
He says: “The history of technology in financial services is about monolithic systems. In banks or pension funds, everything is often done in one place. When technologies change, it becomes hard to keep up to date because every function is wrapped up in a core system. What we are proposing to clients is the concept of a digital ecosystem, where different bits of best of breed software from different companies are used. You end up with smaller tools doing a really good job of discrete functions.”