Supporting the underwriting process

A slick process is crucial to any successful business but to what degree can underwriting manuals fit in with key processes and how can they integrate with other systems such as automatic underwriting engines, workbenches, knowledge management and back office systems?

How does the manual tie in with any automatic underwriting system? Well for sure the automated system needs to reflect the current philosophy contained in the manuals and this needs to be properly coordinated. But sometimes it is not that easy as the manual may be changed with little notice. In SelectX’s rules engine survey of 2011 over half of respondents noted that they were given less than three months’ notice of any changes to philosophy by their reinsurer. So the tools for maintaining the rules engine need to be good enough to cope with this.

Worksheets and calculators

Underwriters need some sort of work sheet on which to record salient points about a case and how they reached their underwriting decision. Some do this on paper and some use an electronic version which is part of their new business system or workbench. A number of underwriting manuals also provide this technology, often with the ability to ‘post’ ratings to the worksheet in much the same way as you populate an online shopping cart.

Last year SelectX 2011 undertook an underwriting manual usage survey to which over 200 respondents worldwide from every major insurance market responded. We learned that around half of the manuals being used have some sort of online worksheet, although of the respondents whose manuals contain a worksheet only a third actually use it.

Some worksheets have been designed to assist with the submission of facultative cases, either by using the worksheet to summarize the case or as part of the covering information when case details are submitted using a secure file transfer site. Alternatively reinsurers have developed either their rules engine, or have built other systems to facilitate this process.

One area to consider when developing a worksheet is whether the worksheet sums the ratings or leaves this to the user. Should there be some sort of algorithm behind the scenes which deals with combination risks? Where rating factors are added together but there is no algorithm is there a risk that the user thinks there is one and so does not engage his or her brain to think about the potential for combination effect? We believe that this risk already exists where some manuals contain a cardiovascular (CV) risk calculator that adds the ratings for individual risks together but does not make any allowance for the sometimes multiplicative interaction between risk factors.

Some CV calculators now are no longer calculators that simply reference values in a couple of look-up tables and add those values together. Some now recognize that BMI and blood pressure are continuous variables, allocating ratings accordingly, and also recognize that knowing or not knowing one variable can impact another (for example blood pressure and lipids). Some employ an algorithm that is quite simple and use simple risk factor scoring, whereas others build more complex mathematical models, often reflecting clinical algorithms derived from epidemiological and other studies.

But does complex mean better, and simple mean worse? Sometimes precision and complexity comes at a price which results in there being no underlying tables to which an underwriter can refer to sanity- check the output of a calculator. Exam students often ask themselves which is more important: the result or the working out?

In 2009 SelectX asked the nine reinsurers then operating in the UK market to underwrite four CV case studies for life coverage. The results were interesting, with decisions on the same cases ranging from standard rates to significant extra mortality ratings. You might think an evidence-based approach would mean that ratings end up the same from one manual to the next. Apparently not given our findings – evidence has to be interpreted, which introduces an element of subjectivity.

From our 2011 manual survey the overwhelming majority of respondents thought that calculators help the underwriting process; a little under ten per cent stated they think they hinder the process. Calculators are certainly on the increase. Our respondents listed tools ranging from the commonplace ones for build and blood pressure to calculators that assist with driving records, renal function and travel risks.

Knowledge management
By knowledge management we mean the systematic capture and sharing of knowledge across an organization so that it can be exploited by employees. Some companies have built their own knowledge management systems to store their knowledge, such as procedural notes, unusual enquiries or their own variations in underwriting philosophy.

Underwriting manuals can be used as part of an organization’s approach to knowledge management and in some cases the manual can be used as a means to index and store information using the existing structure. But if an insurer relies on a manual supplied by a reinsurer, what happens if that reinsurer then takes the manual away in the event of termination of the reinsurance agreement?

Tele-interviews

We have written before about the tele-interview process and manuals. In our recent global survey 90% of respondents said their manual has not been updated to take account of tele-interview information. And over half said that the decision making process for tele-underwritten cases is not supported at all or in respect of less than 10% of cases.

These results are disappointing – although not surprising. Manuals are clearly not keeping pace with new developments in underwriting. Without sensible guidance for the underwriter who is reviewing the information provided by a tele-interview there is a real risk is that the benefits of introducing tele- interviewing, such as reducing the ordering of further medical evidence, is not realized. In fact there is a risk that tele-interviews will actually make the situation worse if more detailed disclosures lead to increased requests for APSs.

Click here to download this report as a PDF