Standing on the precipice

Following the recent Financial System Inquiry (FSI), Australia appears about to embark on a period of innovation in the retirement income sector. If Murray’s recommendations are adopted longevity protection is likely to be a prominent feature. Georgina Hemmings draws insights from the UK annuities market to discuss ways to use mortality and interest rate assumptions to achieve a better outcome for retirees. 

A view from the Summit

In preparation for the journey ahead, retirement income products had a key role at this year’s Actuaries Summit. A common theme from the papers presented appears to be a reliance on average life expectancies not only to simplify modelling but also to underpin product design. Competition in the traditional space has forced insurers to sharpen their pencils, and information and data quality are at the core of differentiated and market leading pricing. Nonetheless, when we talk about managing anti-selection associated with longevity, the same rules do not appear to apply.

Some interesting ideas on group self-annuitisation schemes (GSA) were presented. Together with deferred annuities, this product type is also mentioned in the final FSI report. Setting up a smoker only GSA was discussed during Q&A as an option for addressing anti-selection risk with an assumption that a GSA could not accommodate or, more accurately, attract the business of lives with lowered life expectancies. While listening to this I was jumping up and down in my seat wanting to say we underwrite traditional products, why not longevity?

Enhancing the annuity experience

Mortality assumptions comprise three components which address basis, parameter and trend risk. Translated this means choice of base life table table, the multiplier applied to this table, and the improvement factors which vary mortality factors over time. Taking into consideration an individual’s medical history and lifestyle, underwriting influences the multiplier, but can also impact choice of improvement factors too. Underwriting rewards those in ill health, whose life expectancy is lowered, with higher annuity rates. Without it, insurers rely on cross subsidy and the consequential uncertainty is managed with contingency margins which, ultimately, the customer pays for. For me, cross subsidy isn’t synonymous with risk pooling.

The UK has a well-established underwritten, otherwise called enhanced annuities market. It is suggested over 50% of retirees would be eligible for some form of enhancement[1] and that underwriting, on average, generates 20% higher rates than conventional annuities[2]

In 1995, partnering with several providers, Hannover Re pioneered the development of the underwritten market, which started with a smoker only annuity. RGA was next to enter, in collaboration with other insurers. Today, Swiss Re also provides reinsurance support, and others, such as Partner Re and Pacific Life Re have a slice of the pie. At the same time, leading insurers have sought to develop their own underwriting basis in an attempt to manage costs and improve pricing. 

In addition to age and sex, at the most basic level, providers can differentiate based on postcode or annuity size. Both act as proxies for socio economic group and rely on the ‘wealth health effect’ to estimate life expectancies. These rating factors are particularly useful when gathering medical information is difficult, for example defined benefit de-risking schemes. While postcode and annuity size is readily available, a key challenge of using blunt instruments is the absence of information about core drivers. The European Court of Justice recognised this in 2012 with a ruling that outlaws the use of gender as a rating factor. Even though this was driven by the prohibition of sexual discrimination in insurance, it highlights the need for insurers to understand and seek information on root causes rather than settling for proxies.

The next step up from this is to gather information from applicants about factors which may adversely affect their life expectancy. Major annuity providers in the UK have collaborated to agree a set of standardised questions called the ‘retirement health and lifestyle form’. This form covers personal, lifestyle, and medical questions. It includes supplementary questionnaires for heart conditions; diabetes; cancers; stroke; respiratory and lung disorders; multiple sclerosis; neurological diseases; and where appropriate, activities of daily living. In contrast to the traditional protection market, medical tests and reports are not routinely collected as part of the application process. Instead concerns about over disclosure can be managed by random post acceptance screening. Cotinine tests can be used to substantiate smoking, likewise GP reports for other conditions.[3]   

Underwriting is used to price lifestyle, enhanced, and impaired annuities depending on the level of impairment.

Lifestyle annuities rate lives for smoker status, drinking habits, marital status, BMI, cholesterol, and blood pressure. Some annuity providers have had success targeting this type of business by only underwriting for these factors.    

Enhanced and impaired annuities take into consideration ratings for other illnesses with the delineation depending on severity of impairment. Rating factors for various illnesses and comorbidity are usually provided by a reinsurer’s underwriting manual, although some UK insurers have sufficient credibility in their own experience to develop their own factors. 

There is some interest amongst insurers in varying improvement factors depending on underwriting outcomes. Using a causal based approach to understand the past and predicting the future, the outlook for medical advances in early detection and treatment, in addition to discovery of both cures and new diseases, is likely to vary depending on condition. To demonstrate this, while mortality from circulatory diseases, cancers, respiratory diseases, and external causes (which accounted for 75% of age-standardised deaths in Australia in 2012) have been declining since 1986, mortality from endocrine, nervous system, and mental disorders (accounting for the next 14% of deaths) have been steadily increasing.[4]

The elephant in the room

The importance of interest rate, as well as mortality assumptions, should not be overlooked. A recent study by the UK Financial Conduct Authority (FCA) found that an annuity provider pricing using 2006 interest rate assumptions could offer annuity rates 11 per cent higher than the 2014 level. Similarly, if 2006 mortality improvement assumptions where used, 7 per cent higher rates could be offered to today’s retirees.[5]

In the current low interest rate environment it is challenging to convince retirees to lock in rates prevailing at this time, which is implicit with products that offer a guaranteed income for life. The viability of longevity protection products is dependent on there being a deep and liquid market in which to invest, however availability of long dated bonds is an issue for Australian insurers. For several years the UK enhanced market has relied on equity release assets to bolster annuity rates. Interest rate assumptions on these mortgages are in the vicinity of 150 to 300 bps higher than investment grade corporate bond yields and their long duration can sometime assist with cash flow matching[6]. Similarly there is some interest, and perhaps appetite for, alternative assets, such as infrastructure as a way to improve pricing.

The interaction between underwriting and earnings assumptions should not be overlooked. The role of underwriting becomes even more important in a low interest rate environment as more weight is given to longer duration cash flows once they are discounted. 

Next steps …

Like many of us, I am eagerly awaiting the Government’s response to recommendations from the FSI, particularly around implementation of default comprehensive income products for retirement.

Australia is challenged with encouraging retirees to self-manage longevity risk at the same time as addressing the perception that longevity protection products, in particular annuities, do not represent value for money. A key issue for longevity protection writers is being able to offer a reasonable return to compensate retirees for handing over their savings, sometimes without a return of capital. While product innovation is essential to navigate this, hopefully some of the ideas presented in this article facilitate this journey.


You can contact Georgina at




[4] Cumpston, R., Sarjeant, H. and Service, D. (2015). Estimates of individual life expectancies. Australian Journal of Actuarial Practice, vol 3, 35-46.

[5] Financial Conduct Authority (2014) The value for money of annuities and other retirement income strategies in the UK.


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