Big Data and the digital economy

Over the past decades, data, analytics and artificial intelligence have become increasingly ubiquitous in business operations. In particular, the financial services sector relies on data to understand customer and market behaviours, assess risks, predict outcomes and manage financial uncertainties with aims to deliver efficient outcomes for customers.

Given this context, the Australian Bureau of Statistics and Reserve Bank of Australia collectively hosted the Economic implications of the digital economy conference, which collated perspectives from various experts on the growing digitalisation of the Australian and global economy.

The actuarial profession has traditionally utilised data, statistics, and algorithms to support financial services organisations, and in particular, insurance companies.

The actuarial perspective on Big Data and the Digital Economy: Benefits and Pitfalls in the Insurance Industry represents a critical contribution to the conference, as per the presentation by Elayne Grace. This article provides a summary of the key items therein.

What is ‘Big Data’?

Big Data refers to a greater variety of data, which are available with increasing volumes and are communicated with greater velocity. These may come from new data sources such as telematics and smart devices, satellites and private financial transactions, and have been leveraged with the assistance of continued developments in machine learning and artificial intelligence (AI) programming.

Unstructured datasets and increasing size and complexity rely on such non-traditional techniques to uncover complex patterns, at faster and more granular levels, to assist with decision-making.

In order to exploit the benefits of machine learning and AI, there ought to be adequate considerations of societal attitudes, ethics and risks arising from digitalisation. More transparency and explanations on the use of AI will also benefit consumers interactions with financial services.

Managing Big Data

Given the increasing complexity and volume of data becoming available, it is essential to consider what can be collected and used. There ought to be greater safeguards in place to support enhanced access to public and private data, including intellectual property ownership and the impact of data sharing on societal wellbeing.

In this regard, the Consumer Data Right (CDR) was introduced in November 2017 which gives consumers the right to share their data which whom they elect. The CDR is intended to be progressively rolled out across many sectors of the economy and could therefore have significant implications.

Furthermore, there ought to be a clear legal framework which provides guidance and sets boundaries on the collection and use of data, taking into account fundamental issues of privacy, fairness and anti-discrimination.

Implications of Big Data and the pricing of risks

With the greater availability of granular data, service providers have access to a ‘potentially’ better understanding of their customers’ risk profiles, and this has several implications on financial products. Using insurance as an example, more data will usually lead to greater accuracy in identifying customer risks and putting an estimate on the risk to enable the insurance company to revise the pricing on products. From a modelling perspective, this may lead to a dispersion away from the mean value towards the ‘true’ distribution of risks.

In doing so, the insurer may rely on risk signalling through advancements in connected services and the Internet of Things (IoT) to be notified of customer behaviours in real time.

For instance, telematics in motor insurance allows for the tracking of the drivers’ acceleration, braking, location and other details as they drive.

Whilst this may encourage drivers to have greater awareness on the road, there remain questions on how telematics information should be provided as well as potential concerns over privacy. There ought to be clear boundaries of acceptable risk signalling so that society can yield the benefits of risk reduction and associated lower expenditures on premium, with reasonable safeguards on the use of such data.

Another implication of more sophisticated analysis of Big Data is the potential for insurance to be less accessible and/or affordable, especially when risks are beyond the consumer’s reasonable control and the coverage is essential. In such instances, government may have a fundamental role to play.

As an example, a reinsurance pool is being created to alleviate pressures on premiums for cyclone and related flood damage in Northern Australia, which is backed by a $10 billion government guarantee covering residential strata and small business property insurance policies.

It is important that when more data becomes available for analysis, the re-pricing of risks takes into account what is within and beyond the customers’ control and retains incentives for customers to effectively manage risks, as informed by social expectations and legal requirements.

Furthermore, the digital economy has triggered innovative business approaches in insurance. In addition to the use of telematics in motor insurance, the industry has seen the use of incidental insurance which transforms large and fixed overhead costs into multiple micro-costs incurred on demand which can be easily invoice. As Fintech and Insurtech continue to develop, platforms-as-a-service are also becoming more prominent, where customers may join or create insurance mutuals and effectively reduce the size of under and uninsured persons.

Conclusion

Evidently, there is a plethora of new opportunities that come with the continued advances in data, analytics and artificial intelligence. In order to maximise such opportunities and effectively navigate the challenges that arise therefrom, there ought to be a clear legal and regulatory framework. This will provide guidance on the conduct of practitioners on the access and use of data, and establish what is acceptable within the boundaries of societal norms and expectations.

CPD: Actuaries Institute Members can claim two CPD points for every hour of reading articles on Actuaries Digital.