It is well known within auto insurance circles that there has been a significant rise in high speed auto accidents since Covid. This has led to a spike in auto fatalities.

What hasn’t been explained well is why. There have been plenty of theories about fewer drivers on the road leading to faster speeds or more people on the road outside rush hours.

But none of these explanations quite felt right. There was clearly something missing. Some tried to ascribe it to increased marijuana use. That likely is having some impact, but it doesn’t seem enough to explain the magnitude of the increase.

The Missing Link

Recently, I came across this article in Insurance Journal and suddenly everything made sense. Speeding tickets have gone down since Covid. While correlation does not imply causation, I’m going to go out on a limb and say fewer speeding tickets have caused more accidents.

How is it that insurers have not cited this as a cause of increased fatal accidents? I don’t know. It’s odd. This data comes from TransUnion and I suspect most insurers subscribe to it.

They have made plain there has been a bias towards more high speed accidents. The biggest deterrent to driving at excessive speed is the risk of getting pulled over. If that isn’t there, then, of course, accidents will rise.

The Unconstant Constant

I suspect they didn’t find it because they didn’t look at it. Historically, police enforcement was likely fairly constant and therefore not a variable worth measuring.

Think of it like the constant in a regression model. That constant picks up a number of factors that don’t vary with the result. Actuaries never think about them.

But what if suddenly one of the factors in the constant had a shock and changed? Your model wouldn’t work so well anymore.

So insurers could observe there were more high speed accidents but used their traditional variables (miles driven, rush hour traffic, etc.) to try to explain them to no avail.

Apparently, nobody looked for relationships that used to be stable and changed.

Free Riding

I’ll admit, I never though about law enforcement as an influence on frequency before. It’s so obvious, but it gets taken for granted.

Maybe insurers didn’t want to admit out loud that they free ride off the efforts of law enforcement? But the truth is they are really just a transmission mechanism.

If there are more speeding tickets and fewer accidents, insurers can pass along lower rates to taxpaying drivers. If there are fewer tickets and more accidents, they raise rates to taxpayers.

So I don’t think there is any reason to suppress the relationship. It’s much more likely that it was a “black swan” they had never considered before.

Pandemic Impact

Clearly, Covid was a big driver of decreased traffic stops. I’m sure officers didn’t want to be in close contact with drivers during the initial pandemic phase. There is a huge drop in speeding tickets (and miles driven) over the first half of 2020.

A fair amount of this recovered though in second half 2020 and certainly some of the ongoing effect could be attributed to reluctance to engage with drivers during local Covid outbreaks.

However, the trend has continued to decline over the last two years suggesting there is a bigger impact than just Covid.

Policing Behavior

Without engaging in any political commentary, factually, we can say there have been changes in police behavior over the last few years.

For example, staffing is down. Perhaps with more limited resources, departments are pulling cops off of traffic duty and redeploying them towards preventing violent crime?

It may also be risk management by police chiefs realizing traffic stops create higher risk of confrontations. Or perhaps officers let more speeders go by because they don’t want to be put in a situation where they can look bad on the body camera?

Whatever the reason, it seems likely that these reduced traffic stops are a deliberate strategy by police departments. Thus, it is unclear what will change to make it revert to normal in the near term.

Unintended Consequences

Regardless of your political view towards policing, it is clear that nobody thought of secondary impacts from fewer traffic stops.

Over 5000 more Americans are dying each year in auto accidents. Clearly, not all of those are due to fewer highway troopers, but some percentage of those are.

Part of that increased cost is being borne by auto insurers in the form of higher claims.

Normally, when we think of auto insurers having political risk it’s in regards to activist insurance departments (which obviously remains an issue, particularly in California) but the idea that political behavior impacts auto frequency is clearly a new one and insurers have been caught unaware.

Compounding Impact

Worse, this spike in frequency comes at the same time as insurers have experienced well known severity pressures such as higher used car prices, shortages of parts and labor, and increased lawyer representation.

Thus, the increased frequency and greater damageability (from higher speed crashes) has come at the absolute worst time.

Yet, investor attention has focused primarily on the severity pressures and when they might end. Part of this is surely because it’s easier to look at a Mannheim index every month and assume it is predictive of when better times will come.

Getting Ahead of the Curve

But these severity pressures have been known for the last year and more. Yet, insurers continue to be behind the curve on pricing. That suggests something else is amiss and the continued frequency pressure is probably a big part of the explanation.

Insurers (and investors) need to start watching data on police activity. I don’t want to be a commercial for TransUnion so perhaps there are state sites where this data is also available.

Regardless, insurers probably need to assume higher frequency until we see a return to normal levels of ticket revenue. And maybe they should direct some of their lobbying activity to getting more patrol cars on the highway.