Study Explores Auto Premium Leakage

• Posted by Admin on February 11, 2010
• 2 comments

In January an insurance consulting group out of San Francisco and owned by ISO named Quality Planning released a report about private passenger auto premium leakage.  I will not attempt to make this article a book report about their findings, but rather will comment about some of the more salient aspects of those findings.

The report is based on actual audits performed by Quality Planning in 2008 covering in excess of 4 million auto policies from carriers ranging from substandard to preferred.  Industry estimates were extrapolated from the findings of those audits.  Error margins or degrees of confidence were not discussed in the report, so I will simply state that up front, along with the fact that any numbers referenced in this article are quoted directly from the report with that caveat.

The first major disclosure of the report is that their total estimate for premium leakage for the entire PPA industry in 2008 was $15.9 billion.  This compares to a total premium for the year of $164.25 billion (i.e., would increase the industry total by 9.7%), and would rank as a strong #3 writer if it represented the stand-alone premium of a single insurance group.

As I said, I don’t intend for this to be a book report, so I will just touch on the more interesting findings.  The recommended solutions are pretty much common sense, and are what Quality Planning is selling, so I will skip over those.  At the end of this article, though, I will provide a link for those who wish to see the full report.  Here are some observations:

• In recent years profit for many companies as a % of premium has been depressed.  However, even at a 90% combined ratio, if leakage of 1% of premium can be identified and fixed, that equates to a 10% lift in underwriting profit.  For most companies, the lift would be significantly greater.
• Under the heading of risk management, the report discloses a 200% loss ratio for undisclosed 16-year-old male drivers in their study.
• A fairly obvious observation of the study is that unreported rating information correlates with higher loss ratios.  No surprise there.  But carrying that concept to its conclusion, that means the honest subsidize the dishonest (again obviously).  While the majority of agents and insureds may be honest, a minority gaming the system can throw premiums and losses out of equilibrium, which in turn affects the rates for all.  This is because, what is not so obvious, it only takes a few to mess it up for the many.
• Rating errors that lead to leakage are not random.  Many are intended to lower premiums of otherwise high-rated insureds, introducing a bias into company statistics, whether actuarial, underwriting, claims, or whatever.  Management can be misled as a result.  Here’s where running your book through a tool such as Quadrant’s Quote Converter can have a side benefit in identifying the emergence of unexpected rating cells, besides helping with your competitive analysis.
• Companies are not blameless in the leakage department.  Besides the obvious upfront lack of diligence that can contribute to leakage, today’s ever more complex rating schemes, and the constant rating changes associated with them, mean that insureds are often rerated at renewal, creating more opportunities for error, missed details, or convenient mental lapses.
• It certainly does not help things that certain individuals take advantage of legitimate internet sites suggesting how to lower one’s premium, and put their advice to illegitimate use.  A well informed fraudster is an effective fraudster.  I didn’t spend much time looking, but I didn’t find any sites telling you how to cheat, although I found at least one questionable chat room (but it was aimed at gaming life insurance rates for a smoker, unrelated to this specific discussion, but symptomatic nonetheless).
• Dynamic lifestyles combine with dynamic rating schemes to result in many more opportunities for good old fashioned rating errors (i.e., unbiased) than ever before, in addition to the intentional stuff.  An example – is employment used in rating?  Most workers probably don’t know who Ozzie and Harriet are any more, let alone live like them.  But workers move around now – how current is the employment-related rating information on insureds for whom that is a variable?  What system is in place to keep the information current?
• I won’t get into the specific sub-categories here, but looking only at broad categories, of the above $15.9 billion of leakage:
        - $6.5 billion (4.0% of premium) were attributable to vehicle-based rating errors;
        - $8.9 billion (5.5% of premium) were attributable to driver-based rating errors;
        - $0.5 billion (0.3% of premium) were attributable to other rating errors.
• The above percentages are averages.  Actual results by carrier vary greatly around these averages based on the programs they have in place to address rating integrity.
• Changes between the 2007 and 2008 studies appear consistent with changes in the economy.
• Loss of personal contact with policyholders via the internet and attempts by companies to automate as much as possible likely have combined with more complex rating plans to add to the growth in leakage.
• Effective programs to address leakage can certainly give a lift to a company’s bottom line (maybe a material one), and may even help them establish better communication with their policyholders (which may not be a positive to a select few).