Following the release of risk calculator Ubble, Munich Re's Lee Lovett provides an overview of the tool and what the industry could learn from it.
Ubble is a new risk calculator that has recently launched and claims to predict the probability of death within the next five years.
The tool is designed for those who are aged between 40 and 70 and simply asks you to answer between 11 and 13 questions depending on whether you are male or female.
As well as the risk calculator itself, the website looks at the relative reliability of various pieces of evidence that were considered when producing Ubble.
This will be of interest to underwriters and actuaries alike as they compare this to traditional life insurance risk assessment.
A gimmick?
So is this tool a bit of a gimmick? Certainly not, it uses a very large UK database and analyses numerous variables to assess how closely they are associated with mortality risk; given the extent of the database the tool should certainly represent the UK population as a whole.
In many respects it works a bit like an automated underwriting rules engine that we see in use at most of the UK life insurers.
The main difference is that the outcome is a probability of death within five years, rather than a formal underwriting decision.
Limitations
Given the predictive capabilities of Ubble, does this mean that we will no longer need underwriters or actuaries to help us assess and price mortality risks?
The simple answer is no as, certainly at this stage, the tool does have some limitations.
The first, and most obvious, limitation is that it only looks forward five years - most life insurance policies will run for much longer, typically 20-25 years and hence Ubble does not really work for assessing these longer term risks.
Secondly, there is some subjective assessment within the questions asked, for example a question for both genders that asks for a personal assessment of your own state of health - given the fairly obvious impact on the outcome (especially if the outcome were cheaper life cover) it's easy to see a possible weakness in this type of question.
Having said that, other analyses have shown that the reliability of this type of question can be tested against various other answers given and it can be more accurate than might be expected.
Underwriter concerns
For underwriters looking at the tool, they may be wondering how the assessment can be made without asking a direct question about height and weight to understand the applicant's BMI (as with all life insurance applications).
As we know, a high (or low) BMI can represent a risk factor, albeit indirectly, for a whole range of conditions including hypertension, diabetes, myocardial infarction, other heart problems and cancer.
The simple answer is that the Ubble tool asks some direct questions about these secondary diseases and hence there is no need for BMI details, especially given the short term survival prognosis.
Gender
In terms of the questions asked, the Ubble tool takes a different approach for males and females, with some material differences between the respective question sets, although around half the questions asked are the same for both genders.
A number of the questions are traditional life insurance style questions regarding the individual's health, however some of the questions are slightly surprising if comparing to life insurance applications.
For example, male lives are asked how many cars they own and drive, alongside details of how many people live in the household and the number of relatives/others living with you; for female lives these questions are replaced by one question asking how many children you have.
Another surprising question asked of both genders is your relative speed of walking and the analysis of the variables in the Ubble study shows this is a surprisingly accurate indicator of mortality.
Perhaps this will provide some food for thought for life insurers as they continually seek to fine-tune and improve their underwriting processes and the use of relevant information for risk assessment.
Given the nature of some of the questions asked in the Ubble tool, a pertinent question is whether some insurers may already hold this data - for example details of people living in a particular household or numbers of cars owned should be readily available.
So does this provide some "predictive" data that could give insurers a head start in assessing risks? Well maybe, but the fact remains that the health related information is still the most reliable and accurate in assessing mortality and so will still need to be requested.
In summary, this is an interesting development that has a certain novelty value. The tool is based on a very credible analysis of a large UK database and may provide some interesting insights that insurers may wish to consider.
If the next Ubble update looks at survival probabilities for 10 years or more then this will suddenly start to look far more significant. Watch this space.
Lee Lovett is head of business development at Munich Re UK & Ireland Life
Further reading
