Stats NZ update data collection approach relating to sex and gender

Stats NZ has revealed that after conducting an extensive public consultation there will be a change to the statistical standard relating to how gender, sex and variations of sex characteristics data is collected and reported. The new standard will ensure that definitions and measures are consistent and that they are inclusive of the transgender and intersex population. Stats NZ has also revealed that the collection and reporting approach is based around a human rights approach.

“An updated statistical standard will inform how agencies collect and report information on gender, sex, and variations of sex characteristics, Stats NZ said today. 

The refreshed standard makes definitions and measures consistent, provides guidance for collecting transgender and intersex population data, and is grounded in a human rights approach. 

“It’s important we collect data in an inclusive way, and our process for developing the updated standard reflects this. The refresh has involved extensive public consultation, input from government agencies, international peers, and the support of subject matter experts,’’ Government Statistician and Chief Executive Mark Sowden said.” 

Advisers and insurers also collect sex and gender information. It would be good to see the same standard applied in order to allow data sets to be compared effectively. A graphic from the Statistics NZ guide is shown below to illustrate how to ask the relevant questions. It seems that for the purposes of insurance data collection the approach recommended is to ask sex as assigned at birth and also then to ask gender (as shown in the third part of step three). When underwriting cover, however, identification of intersex variations would appear to be important. Moving these from the health questionnaire to the part of the application where sex and gender questions are asked would help some respondents a great deal. This is illustrated by the additional questions suggested in step three below.  Statistics NZ Guide to collecting gender sex and variations 2021-04-29 143507
 

 
Visit our website to read this news story and the updated standard:


Growth in life expectancy slows - but there is plenty of room for improvement

New Zealand has relatively good life expectancy (compared to many OECD countries) but still has many opportunities to improve - estimate by our data scientist, Ed Foster, using the major causes of death occurring between age 16 and 65 show that:

If we assume there are factors which are influenceable in bringing New Zealand’s mortality rates down to that of the average of the OECD, we can say that 254 deaths could be prevented annually with 87% coming from the female population.

A huge number of those lives that could be saved are women who die from breast cancer. That's another reason why cancer care and access to non-Pharmac drugs is so important. It is also a good reason why real world data should be the underpinning for insurance product rating. 

Turning our attention to the gap to the best performing country for each of the 10 causes of death, we can see that 2,049 lives could be saved annually but now with the majority (53%) coming from the male population.

That shows that although life expectancy growth has slowed recently, see media release below from Statistics New Zealand, there remains plenty of opportunity for us to improve.  A major contributor in this larger number is road safety. Another major contributor is self-harm. Subscribers to our quarterly life and health report have access to the full analysis. 

Growth in life expectancy slows – Media release

20 April 2021

Life expectancy continues to increase, although the change over time has slowed, Stats NZ said today.

Life expectancy at birth for the population as a whole is 80.0 years for males, and 83.5 years for females, based on death rates in 2017–2019. Life expectancy for males has increased by 0.5 years since 2012–2014, and by 2.0 years since 2005–2007. Life expectancy for females has increased by 0.3 years and 1.3 years over the same time periods.

“While life expectancy is still increasing, the increase over the last few years is smaller than in the past,” population estimates and projections manager Hamish Slack said.

Visit our website to read this news story, information release, and methods paper, and to download CSV files:


Merely stating facts is not enough

In research covering more than 6,000 claims for trauma conditions across greater than 2.6 million policy years, recorded claims causes show that cancer accounted for more than 40% of male claims and more than 70% of female claims. That's a huge share. It astonishes me that claims cause was not recorded for over 1,500 claims - but this gap in the data is more likely to be due to poor /legacy management information systems, than actually paying claims without a cause, it is unlikely to affect the ratio of claims causes. 

Consider another pair of facts: in a 30 year period a male non-smoker may have about a 16% (or one in six) chance of claiming on their trauma policy. Trauma claims enjoy a high claim payment rate - it varies, but in the UK a figure of greater than 90% is common. Now consider how they interact: there is about a 1.6% chance that this person will be unable to make a claim. Trauma insurance is a good bet. 

Clients, living their lives, have little or no idea about the risks and odds. It is up to someone to tell them. What's more, if you are basing product selection decisions on long lists of things that have little or no bearing on whether a claim will be payable then the information is true, but of limited use. Weighting the features by claims likelihood is essential to helping the client make an informed decision. 


If only we could do a better job at preventing brain injuries

Recently our data scientist Ed Foster produced a great paper for us on how many lives could be saved if we were able to level up our performance in the top ten causes of death to certain benchmark countries in the OECD. It is astonishing: up to about 2,500 early deaths could be prevented each year - and the knowledge to enable that is already available and being used in practice in countries like ours. 

But this article really caught my attention. How much crime do you think may be related to mental problems associated with brain injuries? It turns out - quite a lot. Imagine how much misery could be prevented, from mere theft to violent assaults, if we could just prevent more brain injuries from happening, or treat them better when they do. 


Survival curves

Max Roser at Our World in Data posted this great image of survival curves as a great way to visualise increased life expectancy added over the last 150+ years. A similar presentation can also be used to illustrate the probability of survival when demonstrating the issues of mortality risk in decumulation planning for retirement. Our World In Data is a fabulous resource for all sorts of issues, but especially good for mortality and morbidity statistics and ideas for good data visualisation. 

Exz5I1GXMActBgK


Why search is a critical function in product analysis

Search is the latest in developments in information sharing. Seth Godin has an excellent post about the development of information sharing and how that has been a revolution driving forward our development as a species. Life and health insurance research is, I admit, not quite as exciting as the development of written language. However we took a big decision to make all our data searchable.

That means that for almost tens years of product data we can find the product version, the specific clauses, and find the sub clauses and conditions that contribute to a better or worse wording. We quantify the impact of that and our reasons why. It is a trail of data that is roughly 50,000 data points wide by ten years long. Searching across the data enables analysis of how product improvements have been applied by companies over time. We can directly relate those to the price charged over time as well, because we took the same decision for pricing data. Only that data is a huge motorway by comparison. Our most recent pricing database has two million data points and the path goes back almost 20 years now.

To add to these great rivers of product and price data we have some tributary streams. We have tracked commissions data for about ten years, quarterly. We recently took our non-searchable news database and 18 months ago renovated that to make it so that it is properly indexed and can be connected to the product, price, and commission data flows. Now when someone comes to us and says 'how did X product compare on Y date?' we can tell you. We can tell you each of the steps between that comparison and today. It's a great resource for product management, heads of sale, and heads of marketing. It is great for advisers too. Predictions about the future usually use past data extensively. Helping advisers with complaints and challenges, assisting in the analysis for current clients with older products, all that data is available to you.

If you subscribe. 


Asteron Life SME insurance index 2020 findings, and more daily news

Asteron Life’s insurance index 2020 provided insight into the insurance-related decisions made by SMEs. The index illustrated that the number of SMEs with multiple life insurance cover decreased when compared to the findings of 2019, with 38% of SMEs having only one life covered in 2020 and 32% of SMEs having only one life covered in 2019. The index highlighted a decrease in SMEs seeking advice from financial advisers and insurance companies and an increase in independent navigation. The findings of the index conclude that SMEs that received advice were more likely to have a broader range of cover. Of the SMEs that sought advice from advisers, 89%  had life insurance cover, 50% had IP cover, 66% had trauma, illness, cancer cover and 50% had TPD cover. Of the SMEs that didn’t receive advice, 88% had life insurance cover, 27% had IP cover, 43% had trauma, illness, cancer cover and 27% had TPD cover. Click here to see all findings of the 2020 index

 

In other news 

Russell’s piece in Good returns: Data versus human

FMA: Insurance Business "ANZ admits to misleading customers about credit card insurance"