Better ways to gather and analyse information are increasing the accuracy of measuring insurance risk, but Charles Pollack warns this still doesn’t mean customers will automatically be charged for the risk
An adage exists in insurance circles that “there is no such thing as a bad risk, only a bad price”. Therefore it is easy to understand that some insurance companies see excellence in risk pricing as an opportunity to obtain a competitive advantage – if you know the risks better than your competitors then you should have superior profit.
In some cases, insurers are collecting more information from customers. As well as internal efforts, expert bodies have helped the industry greatly increase its knowledge about natural hazards such as flooding, bushfires, storms and cyclones. With improvements in technology and increasing data volumes, insurers – especially the large ones – are able to analyse risk to more sophisticated levels than ever before.
Consider locational risk for instance. The location of the risk is one of the key determinants of premium, particularly in property classes. Otherwise known as ‘zoning’, this has evolved from a simple ‘metro/country’ approach to a much more sophisticated approach, were some insurers are even differentiating within suburbs – particularly for things such as flood risk.
An interesting characteristic of this pricing evolution is that while in many cases the true locational risk is largely unchanged, the understanding of that risk is improving. In other words, it is not that a particular high-risk area is necessarily worse than it used to be, rather it is now better understood and also better able to be identified/justified in being higher rated.
A good example is found in the risk of flood. In many cases the flood risk may have been suspected by local residents and councils, however insurers either did not cover the risk or did so but priced it via a cross subsidy whereby customers not at risk subsidised those at risk. These subsidies were often used in ignorance or necessity – better information simply was not available or administration systems could not distinguish the risks.
An important characteristic of locational risk is that it is immutable. That is, the features that make the location a high risk can not be reduced in the short-term. In most cases, for a customer to reduce their risk (and hence their premium) they have to change location.
In the case of flood, the risk is also correlated with other socioeconomic characteristics, such as income. In NSW and Queensland, the states with the largest number of addresses at risk of flood, low income areas contain a higher proportion of addresses with a high flood risk than the overall average.
Currently most insurers do not include flood as a standard cover in their policies. However, with the move in the industry to cover this as standard, pockets of un-affordability and hence chronic under- or non-insurance will potentially develop. This has impacts for governments who could be increasingly called upon to provide disaster funding in these areas.
Insurers that are slow to improve their pricing in this area, including those that don’t change their wordings, may find a migration of very high risk ‘refugees’ to their portfolio. When events occur, they are likely to see above average claims experience in their portfolio or significant negative publicity and pressure to pay out claims ex-gratia.
Leaving flood now and instead looking at cyclone risk, we find that councils in northern Queensland are getting far more sophisticated in their understanding of storm surge that is coincident with high intensity cyclones. Their main focus has been on evacuation, however insurers are not ignoring this information. Looking at Cairns we see that 20 per cent of residential addresses in the Cairns postcode are at risk of significant storm surge in a one-in-200-year event.
So what would happen if insurers all recognised this risk and priced for it or ‘red zoned’ the area? It is quite likely that the local members of parliament – both state and federal – would become involved. Could there be a move to introduce pricing regulation?
Alternatively, the government could introduce a facility to cover certain types of risk, thereby unbundling insurance and removing some risk and potential return from the insurance industry and its shareholders. In the worst case they could choose to fund this facility with a further impost on the insurance industry (similar to fire services levies).
For the company that is the first to move in this direction there is also potential pain. Taking the Cairns example, distribution outlets or intermediaries in the local area are likely to be very unhappy and vocal if they are uncompetitive for the majority of customers that walk in their door.
At the other end of the spectrum, small market players that do not have the capacity to develop their rating to this level could find their reinsurance costs increasing. Alternatively, they may see erosion of their capital base through numerous losses below their reinsurance retention from events where the losses are well beyond their overall market share.
So what do you do with this enhanced knowledge of risk? One option is to monitor the portfolio mix, rather than simply price for the risk. The portfolio performance will only deteriorate when the mix of business shifts towards the higher risk group. Therefore if you know about these risks you can respond accordingly if they start to migrate your way.
While we have only covered a couple of aspects of this here, the concept extends across multiple perils and rating factors. Monitoring mix across all of these factors and figuring out the net impact can be quite daunting. A solution to this is to maintain a risk index. This can be thought of as a ‘technical best’ set of rates.
The risk index is then compared to a premium index (which is the average charged premium indexed at a common point to the risk index). If the risk index increases at a faster rate than the premium index, there is a problem that needs addressing.
Of course, the prerequisite is that the risk is understood. The chart below shows an example of how this might look in practice, applied to new business written in each week.
In summary, it can be said that developments in technology and information mean that understanding of insurance risk is developing at a rapid rate. This doesn’t automatically mean that this risk should be charged to customers. Indeed, the implications of doing so are complex and far reaching.
However if you know about the risk, at the very least you can monitor its presence in your portfolio. Through the circulation of these monitoring reports in-house, the discussions can also begin on what the right response is for your company. You will then be well equipped to respond, should things deteriorate.
Charles Pollack is principal actuary, home insurance at Suncorp
This is an edited version of a paper given at the Institute of Actuaries of Australia General Insurance Pricing Seminar held in Sydney on 30 November
News