Modelling and management of longevity risk

The analysis of national mortality trends is critically dependent on the quality of the population, exposure, and deaths data that underpin death rates.

The great majority of studies that are concerned with modelling build on the assumption that the underlying population data (typically deaths and exposures) are accurate, but errors in population data do occur and affect a range of important calculations, forecasts and assessments. 

Improving long term financial risk modelling is crucial for various stakeholders, including life insurers, and Professor Andrew Cairns, in collaboration with Professor David Blake of City University and Professor Kevin Dowd of Durham University, has spearheaded a number of innovative stochastic mortality models to deliver impact in this area.

Together, the team developed a range of methodologies which can be used as an early warning system for the detection of emerging anomalies in population, exposures and deaths data.

Saving UK pension funds between £330million and £1billion

The impact of the team's research, and resultant methodology framework, has been significant across organisations in the UK, USA and France. These include Prudential Retirement in the US; the Continuous Mortality Investigation (CMI) of the Institute and Faculty of Actuaries; actuarial consultancies advising insurers and pension funds; UK pension funds; insurers and reinsurers; and the Office for National Statistics (ONS).

The team's work has resulted in reduced prices for the transfer of pension liabilities, saving UK pension funds between £330million and £1billion; changes to national high-age population data; changes to the mortality tables used by actuaries in the US, France and the UK; and changes in the methodology underpinning the UK actuaries' CMI mortality projection tables.

Using the research and new modelling Prudential chose to revise their mortality tables and reduce prices, and have also used the results in all of their UK transactions. The CMI used the methodologies and results extensively in two key work streams – mortality projections and high-age mortality. The research was also a key driver for the Office for National Statistics methodological review of official high-age population estimates.

New cohort–births–deaths (CBD) exposures methodology

A key discovery by the team was that an uneven pattern of births within a given calendar year is a major cause of anomalies in population and exposures data decades later.  Different countries or agencies may derive population and exposures in different ways, but however they do this, the estimates will be subject to potentially significant errors, unless irregular patterns in monthly or quarterly births data are taken into account.

The team's new methodology, called the cohort–births–deaths (CBD) exposures methodology, calculates exposures data by using births data at the monthly or quarterly frequency, and can also be used to improve the calculation of mid-year population estimates based on census data. 

Analysis using the new methodologies demonstrated that the ONS population data contains significant anomalies of up to 9% that need correction. The team's discoveries, analysis and research led to the approach from the US-based multinational insurer Prudential which asked them to investigate these issues further, and to propose how anomalies in population and mortality data could be corrected.