PwC’s Bryan Joseph on how big data will reshape the industry

Data, software

Not a day goes past without the term big data being debated in the media, with lots of comment on how it will transform our businesses and the major impact that it is having on our lives, writes PwC partner and global actuarial leader Bryan Joseph

By contrast, the reinsurance industry seems to be more engaged with the impact of alternative capital, changes to terms and conditions as well as the direction of rates. It is therefore worthwhile asking the questions of whether big data should have an impact on the reinsurance industry, and, if so, what opportunities and threats it poses for us in our operations.

The manipulation and analysis of data is nothing new to the insurance industry, so what is different with so called ‘big data’? There is no universally accepted definition of the term. Schlonberger and Cole describe big data as the “things one can do at a large scale that cannot be done at a smaller one, to extract new insights and create new forms of value”. In essence, it is referring to the fact that the amount of data available and stored on computers, in structured or unstructured formats, has become so large that the quantity that could be examined required the development of special tools to manipulate, analyse and handle it.  The change, though, is not about the manipulation, it is about the analysis and interpretation.

Reinsurers have always had to cope with incomplete or partial data from their clients, and to be unable to see or understand the complete picture.  Analysis had consistently assumed that data will be precise and accurate, but made some allowance for the fuzzy and incomplete nature of the data made available to them. Looking forward, in the era of big data those excuses will increasingly become a thing of the past. Reinsurers would not only increasingly demand the data relied on by their clients, but also combine this with secondary data that will supplement the primary data and enhance their results. Furthermore, as the reinsurers access increasingly granular data, they will begin to face similar challenges that are being, and will be, faced by their direct counterparts. 

Imagine, for example, a reinsurer writing excess of loss capacity on the product liability of a large pharmaceutical portfolio. In a big data world, they will be able to tap into the healthcare or distribution systems of national countries and analyse the users and incidence of specific side effects.

While this data may be anonymised, there have been many cases of this type of data being reconstructed, and problems arising. How will reinsurers protect data that they have access to, and manage security breaches?  Similar issues arise in the case of flood and other risks.  It is in the interests of the insurance industry to ensure that proper premiums are charged for the exposed risks. Geocoding has transformed our ability to understand the precise locations of properties, and, combined with granular and digitised mapping data, to exposure to floods of specific sizes.

While flood is often excluded from standard covers, reinsurers should use this data even when called upon to reinsure national schemes. Clearly this could result in penalising specific areas or exposures, and that can lead to accusations of discrimination by the insurance industry. 

Reinsurers would not be able to escape this challenge.  Big data is said to offer the opportunity to engage more with predictive modelling and to better forecast choices based on in depth statistical analysis. It also opens up the perspective that class action suits may emerge from the impact of big data algorithms on the population at large. It is these types of parallel risks that could prove the largest threat to reinsurers.  The agile reinsurer needs to be alert to changing trends within the market and ensure that policy terms and conditions do not leave them exposed.

Wherever there are threats, there are always risks which create opportunities for reinsurers. The changes in sensor and data technology have led to telematics, and will ultimately lead to driverless vehicles.  The scenario where insurance is bundled into car prices and vehicles are leased, not sold, leads to a different scale of policies and potentially a transfer of this class from personal lines to manufacturers and from insurers to reinsurers. The challenge and question will be: what is the shape and size of this risk, and how can it be priced?

Big data will undoubtedly be the thing that will reshape our industry. But it brings reinsurers closer to the data and thus to the operational risk exposure arising from being a business-to-consumer business, rather than just a business-to-business organisation. On this basis alone, big data is not something to be ignored.