The High Cost of Inaccurate Auto Insurance Underwriting

The property and casualty insurance sector doesn’t have a claims problem. It has an underwriting problem.

Although high losses from severe crashes, inflated repair costs and natural disasters have put pressure on the auto insurance sector, the whole point of insurance is paying claims. There’s only a problem when the claims dollars paid aren’t in sync with the premium dollars received. Therefore, it could be argued that the real issue is inaccurate underwriting.

A Decade of Auto Insurance Profitability Challenges

The property and casualty insurance sector has been struggling to stay profitable, and auto insurance underperformance is a huge part of the reason.

Personal auto insurance is struggling. Fitch Ratings says multiple carriers reported auto combined ratios of 110% or worse in 2023, and some carriers saw their financial strength ratings drop due to weaker personal lines performance, including auto.

Likewise, the U.S. commercial auto insurance segment is looking at a combined ratio in excess of 106% for 2023 according to Fitch Ratings. Unfortunately, this isn’t just one bad year. The line has underperformed regularly for more than a decade, with 11 of the last 12 years seeing an underwriting loss. Given this timeline, it’s clear that recent inflation may have made the situation worse, but the true extent of the problem goes much deeper than that.

The Status Quo Isn’t Working

Insurers typically rely on underwriting rules and third-party data integrations to predict the potential for loss. Recent underwriting performance suggests that this approach is no longer working. Below are a few costly implications:

Precision Underwriting Could Turn the Tide

Something needs to change in the auto insurance sector. Insurers can’t control claims, but they can control underwriting. Instead of focusing on underwriting rules and third-party data, insurers can build algorithms based on actual historical experience data to predict each policyholder’s likelihood of loss more accurately.

With more precise underwriting at the policy level, auto insurers can achieve the following:

Improve underwriting performance.

For insurers, improving underwriting performance is a priority, and adopting a more precise, policy-level underwriting process is the most efficient way to meet this goal.

By only quoting adequately-rated drivers, insurers can very quickly influence loss ratios.

Maintain and grow market share.

When insurers shut off segments in pursuit of greater profitability, they also give up significant market share, some of which is good business. The same effect occurs when insurers raise rates for an entire segment; the best policyholders within the segment are likely to seek better rates elsewhere. While measures like this may be seen as the fastest way to influence loss ratios, it’s a case of throwing out the baby with the bathwater. Conversely, precise, policy-level underwriting enables insurers to turn all segments on while precisely selecting and quoting only adequately-rated risks, one policy at a time.

Provide a positive policyholder experience. 

Policy-level underwriting can improve the policyholder experience in two ways. First, decisions are made in less than a second, and the importance of speed for both brokers and policyholders cannot be understated. Second, policy-level underwriting enables good risks to continue enjoying the good rates they deserve.

There’s a lot of talk about digital transformation, but many companies still cling to the old ways of underwriting. The risk landscape has changed, and it’s time for underwriting processes to catch up. Fortunately, there is a smart solution.

Soteris has harnessed the power of machine learning and focused it on a singular point of failure in the typical auto insurance risk selection process to deliver exponentially better results than would otherwise ever be possible. This solution can be affordably implemented within a four-month timeframe alongside existing systems and using existing rate filings. Why wait another day?

4-Month Implementation