It’s no secret that the homeowners insurance business has historically taken a back seat to the other insurance lines when it comes to pricing sophistication. But now, insurance carriers are investing in their homeowners business as they realize this is an area with household growth potential. Furthermore, by focusing on improving their homeowners pricing precision and use of analytics, insurance carriers have the potential to create a distinct competitive advantage.

Below we will uncover the best practices in ratemaking approaches used by North American insurance companies in their homeowners businesses. Based on results from the Earnix/ISO 2014 Homeowners Ratemaking Insurance Survey* we will take a first-hand look at key areas of homeowners pricing as well as discuss the aggregated results from the real-time polling that was held during the ISO/Earnix live webinar.

1) Predictive Modeling
In the 2013 ISO/Earnix Survey, 82% of survey respondents said they used predictive analytics in one or more lines of business. The most common use of predictive analytics was found in the Personal Auto Industry (49%) followed by Homeowners (37%), Commercial auto (32%) and Commercial Property (30%). Respondents were also asked what line of business they would see themselves using predictive analytics in the future and homeowners came out on top. In our latest survey, over half of the survey respondents used predictive modelling for homeowners loss cost development vs 37% compared to the year before – that’s more than a 50 % increase year on year.

Of those that incorporated predictive modelling into their projects – half of the companies were using internal and external resources for developing models, while 40% of the companies were using only internal resources in developing models.







2) By Peril Rating
As more companies continue to adopt predictive analytics as an accepted best practice, many are also building out by-peril rating structures. While modeling homeowners losses by-peril significantly increases the modeling efforts, companies are seeing significant benefit in their overall estimate of loss.







a) Rating Structure for Homeowners Insurance Varies by the Company Size:
The majority (82%) of the survey respondents representing companies with over $500M GWP use a by-peril rating structure, while most (61%) of the survey respondents from smaller companies combine all perils.









b) Incorporating By-Peril Rating into their Rating Structure:
Over a quarter (26%) of the survey respondents that use all perils combined intend to incorporate by- peril rating into their rating structure within the next year, and another 40% plan on doing it in 2-3 years.

c) Winning with By-Peril
By peril rating is working for those who have adopted it. Over the last 6 years from 2007 -2013, ISO has followed 25 insurers using by-peril plans. In 2007 these companies had 28% market share while in 2013 their market share had grown to 34%.

The study also found that companies using by-peril rating, have loss ratios of 4.3 points lower than their competition. While we cannot say that the growth is solely attributed to by-peril rating, we know that by-peril rating goes together with a culture of leveraging predictive analytics; the growth and profitability improvements could be attributed to an overall level of analytical sophistication amongst these companies.







3) Catastrophe Modeling
a) Using Catastrophe Model Output in Ratemaking
Many years ago, one of the only ways homeowner insurers were able to leverage more sophisticated analytics was by utilizing CAT models. Some of the first perils developed decades ago were hurricane and earthquake, while more recently winter storm and severe thunderstorm have been added. The most common way for using CAT models in ratemaking is by simply adjusting the overall rate level based on the output of the model (62-71% for the various catastrophes).

b) Future Use of Catastrophe Model Output in Ratemaking
Most survey respondents who do not currently incorporate catastrophe model outputs into their ratemaking, have no plans for using it in the future. However, severe thunderstorms (45%) followed by winter storms (32%) and fire (26%) are the most common models that insurers plan to build out.







See our follow up blog with more survey results. Feel free to download the full survey report or listen to the on-demand webinar discussing the survey results.

QUESTION: With recent industry attention being directed towards the home product, do you think companies are justified in investing so heavily in homeowners predictive analytics? Is this trend here to stay, or are companies jumping onto another bandwagon? Leave your comments in the comment box below.

* The survey had 99 respondents from mainly the US (90%) but some from Canada (10%) as well. Respondents represented a good mix of company sizes, with three quarters of the survey respondents coming from companies with up to $500M of homeowners insurance gross written premium (GWP), while the remaining quarter of respondents came from companies writing over $500M in premiums.