Big Data: Blessing or Disaster?

Servaas van Bilsen is assistant professor of actuarial science and mathematical finance at Tilburg University (NETSPAR) and the University of Amsterdam. His research interests include pension economics and behavioral finance.

The pension landscape is changing. So-called Defined Benefit (DB) plans are losing popularity, while Defined Contribution (DC) plans are becoming increasingly popular. This trend brings new challenges for plan members and pension providers. In particular, the question of how pension wealth should be invested and withdrawn is now more relevant than ever before. Big Data can play an important role in answering this question. On the other hand, Big Data may undermine collective risk sharing.

All over the world, pension systems are undergoing fundamental change. A new pension system is currently also under debate in the Netherlands. Much remains unclear about the future Dutch pension system. What seems clear, however, is that the new pension contract includes more freedom of choice for the plan members. This allows Big Data to create value by helping people to make better choices regarding their retirement. The underlying idea is that as more useful information becomes available, better pension products can be offered. More relevant information about the following three aspects leads, in principle, to a better pension contract.

Life expectancy plays a major role in the pace at which pension wealth should be withdrawn. For example, a person with low life expectancy usually has a strong preference for consuming a large share of pension wealth during the first years of retirement as compared to a person with high life expectancy. The optimal retirement age depends on the life expectancy as well. Therefore, useful information about an individual’s health may help to achieve a better allocation of consumption over the life cycle. Increasingly more health data is becoming available. Consider for example loyalty cards from supermarkets that generate an enormous amount of data related to the lifestyle of an individual. The question remains whether this privacy-sensitive information may be made available to pension providers. Politics play a crucial role in addressing this question.

Financial situation
In addition, increasingly more data is becoming available about an individual’s financial situation. Examples of financial data include: the amount of freely available wealth and whether or not people own a house. The optimal pension contract depends, among other things, on an individual’s financial situation. A person who already saves a large amount every year may have less need for contributing a substantial part of his or her salary to a pension fund as compared to a person who does not voluntarily save at all. Moreover, Big Data can be used to notify plan members when their financial situation changes. For example, members could receive a message on their mobile phone in case of insufficient pension savings.

Risk preferences
The optimal pension contract also depends on an individual’s willingness to take investment risks. Relevant information about risk preferences can be obtained through surveys or observations. The latter method may perhaps be preferred, as it considers actual (observed) behavior. Big Data makes it possible to determine the degree of risk aversion. For example, the composition of an individual’s investment portfolio provides useful information about the level of preferred security with respect to the pension benefit.

Big Data not only offers benefits. There are also potential threats. For instance, Big Data may undermine collective risk sharing. Consider for example the sharing of mortality risk. Economic theory teaches us that risk sharing can create economic value. However, this value is lost when people with heterogeneous characteristics no longer want to share risks. Big Data may result in a higher predictive value of individual life expectancies. This raises the legitimate question: “To what extent is a person with an unfavourable prognosis of his or her life expectancy willing to share mortality risk with other people?”

The above discussion shows that Big Data offers many opportunities for pension providers. However, the trend towards Big Data is not without risk. Particularly the willingness to share risks is under threat. Time will tell to what extent Big Data will be valuable to pension plan members.