All Actuaries Summit 2026 Papers

The Failure Metric

Read the paper

Author Ilan Leas

This paper explores lessons from life insurance product failures.

Through ten global case studies, three key conclusions were drawn:
1. Failure in life insurance is slow and cumulative, but with a journey that is predictable. It is typically the result of small decisions compounding, driven by pushing boundaries (stretching guarantees, nudging benefits, assuming behaviours and chasing competitiveness) in an attempt to address a genuine consumer need.
2. Most failures are behavioural, structural and human. Actuaries tend to focus on assumptions and the models, but we will inevitably get these wrong. Whilst there are specific micro lessons to take away from each case study, the failures emerge where misaligned incentives creep in, we forget insurance 101 rules, we ignore distribution realities and we can’t foresee future systems or how outcomes will be judged years later.  
3. There are no silver bullets, perfect designs or foolproof rules we can put in place that stop failure or will stop us repeating history. But we can learn from other industries more used to failure, and we can ask sharper, more uncomfortable questions to help decision makers better assess the risks.  

Ten questions were developed from the lessons and are proposed to encourage longer term thinking by decision makers when designing products. They are:

1. If volatility is inevitable, how have we incorporated the impact into the design?  
2. What risks do our customers think they have transferred?
3. How have we set up this product for the successor team/s that will be managing this in the future?
4. What experiments have we run to test behaviours, both at sale and over the life of the product?
5. What specific lessons from our own past product failures or from failures we have witnessed elsewhere are explicitly embedded in the design? 6. How is this product going to be a no brainer for our customers?
7. How have we built in positive surprises to the design?
8. If we assume some information is being hidden from us in our testing, what could it be and how are we incorporating this into the design?
9. What are all the incremental changes to the product since launch and how does this cumulative upgraded design compare to the original in terms of risks?
10. If we fast forward to the balance sheet 10 years from now, where is it sensitive to changes in assumptions?

The goal can never be to eliminate failure which would be unrealistic in a particularly longterm business defined by uncertainty. But to change its shape. Smaller. Earlier. More transparent. The kind that teaches quickly, still protects customers, and ultimately strengthens the system. Few industries have the same depth of experience, the same long-term datasets and cashflows, or the same ability to observe how decisions play out over decades. This is an asset and is a source of competitive advantage if we learn from it.

Sequencing Risk and Asset Allocation

Read the paper

Author Colin R. Grenfell FIA, FIAA, FASFA

Abstract
Driving a car downhill requires different skills from those required to drive the car uphill. 

Similarly, to secure optimum superannuation member benefits in the decumulation stage, different asset allocation strategies are required from those which apply in the accumulation stage.  Why? This is primarily due to sequencing risk.  

Many authors have attempted to define sequencing risk. The definitions are usually reasonably consistent with the definition in this paper.  Some have illustrated it with simplified examples, but few, if any, have quantified it and then measured it based on historical data.  

This research compares the sequencing risk faced by a retiree or an annuitant invested in a balanced portfolio, with that resulting from investing in:  

- Australian shares,  
- international shares or
- a combination of Australian shares and cash.   

Comparisons are also included for Australian direct and listed property, semi-government bonds and inflation-linked government bonds.  In the final section results are presented for ‘real’ sequencing risk allowing for price inflation.

The comparisons use both the Austmod historical data and simulated scenarios. The comparisons reveal that Australian shares combined with cash, offers better sequencing risk protection than traditional balanced portfolios.

This paper and presentation are an update and extension of an Actuaries Digital article with the same title which was published 30 April 2025.  

Subsequently, that article was published on LinkedIn and since then it has experienced sustained and continuing interest from professional readers.