Despite driving well–yielding politely, obeying every stop sign, and racking no tickets or accidents–your auto insurance premiums seem to have steadily risen in recent years. It’s not just you though. Since about 2012, rates to insure vehicles have gone up in all states and for all drivers, even including those with spotless driving records and no claims.
And not by negligible amounts. Since 2012, the consumer price index (CPI) for auto insurance has gone up by 21.5%, compared with a rise in the overall consumer price index of 4.5% over that same five-year period. It’s the largest five-year growth of auto insurance costs since the early 1990s, when costs grew by about 30% between 1989 and 1993.
The Profit Challenge
The insurers aren’t raising rates for the sake of just charging you more–there are rules against that, actually. Instead, the driving force in the upward march in premiums is an auto insurance industry that’s been finding it increasingly difficult to sustain healthy profit margins.
Of the top five insurers, only GEICO and Progressive have managed to maintain profits, and the amounts by which they’re in the black have trended downward over the last seven years. It’s even worse for some other insurers like State Farm. The single largest auto insurer in the country has seen its revenue from premiums rise by 26% since 2010, but also suffer a 35% increase in the cost of covering the cost of claims in the same time period.
Another way to look at these numbers is through the combined loss ratio: the ratio between underwriting losses (as well as all other business expenses) and written premiums. In 2010, the average direct combined loss ratio was 99.7% amongst the nation’s ten largest insurance companies meaning they were just barely making a profit off auto insurance premiums. In 2016, the ratio ballooned to a whopping 107.1% average, meaning the major insurers were losing 7% more than earning last year.
So what’s causing the insurers to lose so much money?
Car insurers say everyone will pay if a New York regulation, aimed at preventing insurers from using a widespread rating practice, makes them change their business model.
New York Governor Andrew Cuomo has fired a big gun at the auto insurance industry, charging that its practice of rating drivers on the basis of education and job status is discriminatory. That’s likely to echo throughout the country as consumer advocates clash with insurers that say they need tools like this to give everyone lower rates.
Cuomo, a liberal Democrat, has proposed a new regulation — to take effect within two months — that would virtually stop insurance companies from using these yardsticks to determine premiums. “This new protection cracks down on this unfair practice that soaks drivers for not having a college degree or a high-paying job,” said Cuomo in a statement.
Insurers said they’ll challenge the proposed regulation, but their options could be limited. “The New York Insurance Association will be commenting on this regulation and encouraging regulators to consider the impact [this] … will have on a market that is currently competitive,” said association Vice President Cassandra Anderson.
This ruling could raise — or lower — car insurance premiums for about 12 million New York drivers. But it could also have a far-reaching impact nationally. The largest car insurance market, California, had already limited most metrics that insurers used in deciding who pays how much and forced them to focus primarily on an applicant’s past driving record.
With New York, which has 150 licensed insurance companies, moving in the same direction, a snowball effect could result. At least 12 other states have already introduced legislation to keep insurers from peeking at drivers’ educational records and other criteria.
“This is just the tip of the iceberg in removing socio-economic factors that harm the market,” said Robert Hunter, the director of insurance for the Consumer Federation of America. “Credit score, home ownership or any gap in coverage when the person has no car — all have to go too.”
But for a lot of Americans, this could have “the opposite effect and result in many consumers paying more for insurance,” said Vice President Kristina Baldwin of the Property Casualty Insurers Association of America.
“Insurers use factors such as education and occupation because they are actuarially proven to be highly predictive of the likelihood of an insurance loss,” said Baldwin. “For example, some insurers have found that teachers, paramedics and librarians represent lower risks and could be in jeopardy of paying more for insurance.”
Insurers point to numerous studies — by insurance regulators — in Florida and Maryland, as well as other states that show nondriving factors such as education, job status and particularly credit rating, are key to how a motorist will perform on the road.
New Jersey, for example, once had one of the highest premium rates in the nation. Then the second-largest U.S. car insurer, GEICO, began selling insurance in the state in 2004 and brought rates down by rating drivers on these factors and others. Consumer groups cried foul, but New Jersey regulators looked at the bottom line and allowed it to continue.
Auto insurance critics say the only fair ranking criteria is driving accidents, DWIs and speeding tickets accumulated by a driver. But insurers argue that while these are important, carriers are insuring against future bad driving, so unemployment or failing in college has to be considered because they make an applicant more risk-prone.
The insurance industry is a catch-22.
On the one hand, it is a customer-facing landscape, and there is marginal room for error due to the scope of the business. On the other, a vast amount of customer data, usually logged manually by humans, must be accounted for and analyzed. According to Experian, when data is entered manually, incomplete or missing data makes up 55 percent of errors, and typos account for another 32 percent — both mistakes that are easy to miss.
The claims handling process defines the relationship between the insurer and the insured, but it is a major operational hurdle if managed incorrectly. And so insurance agents are faced with a dilemma. How can they transmit data in the shortest amount of time with the highest rate of accuracy? Taken a step further, how can they correctly verify claims, including fraudulent ones, under the same pressures?
Artificial intelligence (AI) is emerging as an effective way to both speed up the claims verification process and improve data accuracy.
Moving past the inaccuracies
In a survey from accounting firm EY, global consumers said they trust the insurance industry less than banking, supermarkets, car manufacturing, and online shopping. Much of this distrust stems from claim inaccuracies, which are difficult to avoid when data is input manually. AI technology helps solve this by reducing the amount of manual input.
Despite advances in technology, many claims processes still rely on humans to manage tasks like matching customer information within numerous databases. Insurance organizations that implement AI to take over this process can match the information in half the time. By automating step one, insurance agents can move on to more customer-oriented tasks, like personalizing the customer experience.
And incorrectly typing information isn’t the only source of inaccuracies. Insurance agents struggle with outdated claim systems and poor data quality, factors that make it difficult to properly manage claims. AI technology bypasses these systems to help increase overall accuracy.
Using AI to recommend claims payouts
Claims cycle time is the leading indicator of customer satisfaction, according to a recent J.D. Power & Associates property claims satisfaction study. The study found the average claims cycle can take as little as 11 days, but there’s no reason this cycle can’t be cut further.
Part of the claims cycle is deciding whether or not to actually pay the claim, and this can be time-consuming. By integrating AI technology, insurance companies can use cognitive analysis to quickly recommend the appropriate payout. First, AI mines the information and auto-validates the policy. Then, using machine learning models, it makes decisions on the claim and chooses whether to automatically transmit data into the system for payment. AI has the ability to analyze structured data (i.e., online databases) and unstructured data, like handwritten forms, letters, and certificates, to make a cohesive recommendation. Less than 30 percent of insurance customers globally report having positive customer experiences — and AI technology may be the difference between a good customer experience or a bad one.