Heading into the New Year, personal auto carriers are preparing to start subtly taking rate to account for increasing frequency and severity trends.

Premiums won’t spike—“we don’t anticipate any real significant changes in personal auto rates—maybe a slight uptick, but nothing like what we’re seeing on the commercial side,” says Richard Kerr, CEO of MarketScout.

But factors like higher mileage, distracted driving and more sophisticated auto safety technology are combining to put pressure on the market, confirms Jon Bloom, vice president, personal auto at Erie Insurance, who says his company is “seeing frequency and severity increasing at the same time.”

Although many anticipated a rise in severity based on in-auto technology advancements and increasing medical costs, it’s still an unexpected direction in a market where frequency, at least, was expected to start declining.

“Historically, the personal auto industry has seen a longer-term decline in frequency driven by a combination of things—safer cars, safer roads, better drunk driving laws,” explains Karen Bailo, agency distribution business leader at Progressive. “Based on the long history of negative trends, we would have forecasted a gradual decline in frequency.”

Not so, say the numbers. “From what we’re seeing, the reality is different,” Bailo says. “I think many of our competitors are seeing this as well—in the past few years, frequency started to increase, putting upward pressure on prices.”

But with driverless cars just around the corner, isn’t this hike in frequency likely to be short-lived? Not necessarily, Bailo cautions: “We really think driverless cars are further off than you might expect. That shift is inevitable, but a few key factors that will determine how fast it happens, and whether it accelerates or not.”

Here are four:

1) The current fleet must completely turn over. “People are holding on to their cars longer,” says Bailo, who notes that the average age of a vehicle currently on the road in the U.S. is up to 11.6 years from 9.6 in 2002. A 10-year-old car won’t boast the latest and greatest in terms of autonomous technology.

2) The technology penetration rate must increase substantially. Considering the age of the average U.S. car, it’s no surprise that only about 5% of today’s fleet incorporates driver assist technologies such as automatic braking or lane departure warning, according to Bailo. And it’s not because the technologies are simply too new, she adds—forward collision avoidance, for example, was first introduced about 10 years ago.

That’s a much slower penetration rate than earlier vehicle technologies, Bailo points out: For example, “when electronic stability control was introduced 25 years ago, 10 years later, there was probably about 15% penetration.”

“Even if vehicles today have features of autonomous vehicles, they still rely heavily on humans,” Bloom agrees. “We expect that to continue on for the next several years. Whenever safety advances come into play in vehicles, they’re incremental—they haven’t historically been this giant big bang.”

3) Driver assist technology must move from optional to standard. “Right now, a lot of those safety features and autonomous features are either in constant pilot stage, or they’re optional,” Bailo notes—and as long as that’s true, “you’re getting into an affordability issue” for consumers.

4) Driverless technology must advance significantly. It’s one thing to drive a car that can automatically hit the brakes if it detects you’re about to crash, but “there’s a lot more complexity when you get into higher levels of automation where the driving system becomes more responsible for monitoring the whole environment of the car, as well as where the car is in the surrounding ecosystem,” Bailo points out. “There is going to be a gradual advancement in technology, but it will take some time.”

Eventually, as automation advancements gradually penetrate the fleet, they should “ultimately put some downward pressure on frequency,” Bailo predicts.

Based on fleet penetration rates, Progressive estimates “it could take as long as 40-50 years for a new technology to fully penetrate the fleet,” Bailo explains. “Even if that technology is effective at ultimately reducing accidents, it would take quite some time before the full effect is reflected in our frequency trend.”

In layman’s terms, Bailo says, “we’re pretty far off from full automation to the point where we think it will have substantial impact on frequency in one direction and severity in the other direction.”

Jacquelyn Connelly is IA senior editor.